Final Report Summary - GGP (Generations and Gender Programme: A European Research Infrastructure on the Causes and Consequences of Demographic Developments)
Executive Summary:
The Generations and Gender Programme (GGP) is a Social Science Research Infrastructure that provides micro- and macro-level data which significantly improve the knowledge base for social science and policymaking in Europe and developed countries elsewhere. In the period 2009-2012, a Design Study, funded by the EU’s Seventh Framework Programme has been carried out.
The key features that allow the GGP to make a significant contribution to the scientific and policy challenges facing European and other developed societies are:
• Cross-national comparability;
• A longitudinal design, with a three-year interval between survey waves;
• A large sample size, with on average 9,000 respondents per survey wave per country;
• A broad age range, with respondents aged 18 to 80;
• The combination of micro and macro data;
• A theory-driven and multidisciplinary design.
Nineteen countries have conducted at least one wave of the Generations and Gender Survey, while 12 countries have carried out multiple waves. By March 2013, wave 1 micro-datasets for sixteen countries and wave 2 datasets for six countries have been made available to researchers. These data are used in over 500 research projects and have led to over 600 articles, reports and scientific presentations. The GGP Contextual Database provides information for 60 countries on more than 100 indicators.
The Design Study has been used to evaluate the content and quality of the data collection and dissemination procedures and practices. On the basis of that evaluation, plans for the future have been outlined. The most important of these are:
• The development of a questionnaire for the 2015 wave of the Generations and Gender Survey (GGS), showing a well-balanced mix of continuity and innovation.
• The adoption of a mixed-mode survey design to reduce data collection costs, whilst maintaining scientific integrity.
• Regular updating of the Contextual Database, with a focus on increasing the number of available policy indicators.
• Streamlining of data access procedures, to expediently release downloadable micro-level data.
• Sustained efforts to disseminate GGP results to scientific and policy audiences, through our monthly newsletter, a GGP Research Notes series, and by small- and large-scale policy events.
The Design Study has also been used to work towards a sustainable future for the GGP. This includes:
• Increased institutionalization. The GGP undertakes all possible measures to assure inclusion in the ESFRI Roadmap. In addition, the GGP will work towards the establishment of an ERIC.
• Costs reductions. The GGP is committed to run the research infrastructure as cost-effectively as possible. A series of measures to reduce costs have been and will be implemented.
• Fund-raising. The GGP will continue to support fundraising activities in participating countries, but also seek central funding of data collection activities
Overall, the GGP provides a major research infrastructure to inform scientific research and policy debates on issues like intergenerational solidarity, ageing, the nature of family structures, newly emerging social inequalities, work life balance and low fertility; moreover such debates have to be examined in the context of the fiscal crisis and existing policy and structural constraints.
Project Context and Objectives:
In an increasingly globalized world, Europe faces major social, economic, cultural and technological challenges. In Europe 2020, the European Union develops a strategy “to help us come out stronger from the crisis and turn the EU into a smart, sustainable and inclusive economy delivering high levels of employment, productivity and social cohesion”. To realize the smart, sustainable and inclusive growth that the EU envisages, the human capital of Europe’s populations is the continent’s largest asset. To capitalize on that asset asks for a constant reappraisal of how Europeans arrange their lives, both as individuals and as countries. The economic crisis affects not only day-to-day decisions, but also fundamental choices at all stages of people’s lives: marriage and childbearing, the combination of employment and caring responsibilities for the young and the old, retirement, housing, and ageing well. How Europeans cope with these fundamental choices has important consequences for their personal well-being, but also for the adaptability and competitiveness of the societies in which they live. In addition, current demographic developments in Europe have far-reaching implications across all spheres of society, including economic development, social cohesion, sustainability and equity. These developments pose key challenges to public policy-making. The GGP design study aimed to improve our ability to explain the causes and consequences of demographic developments, in particular the changing relationships between genders and generations.
Population challenges. Major demographic developments like increased longevity, low fertility, and high levels of marital instability pose – in particular in combination with other developments like increased female labour force participation – major challenges to European societies. These challenges have become even more urgent given the economic and financial crisis that has struck Europe – and other Western countries – in recent years. Therefore, understanding these challenges and providing answers to them is important both from the perspective of individuals and families, and from the perspective of societies.
Gender relationships. Individuals and couples have aspirations with regard to parenthood and occupational careers and try to strike a balance between the two. At a societal level, a balanced level of fertility and a sufficiently high level of labour force participation are also important to assure a sustainable future.
Intergenerational relationships. People need care. Families are the most important providers of such care. Support is provided both from the older to the younger generation and vice versa. For families, a major challenge is how to uphold the provision of care in situations of increased longevity, smaller families and decreased time availability due to labour market obligations. At the societal level, a challenge is to prepare for a future in which the demand for care is almost inevitably expected to rise.
To improve our understanding of these major societal changes, and their causes and consequences, and to develop appropriate policy responses, top-quality, comparative data are a necessity. In order for the social sciences to provide answers to these questions, data are needed that (1) cover the complete adult life course and focus on key decisions and transitions in men’s and women’s lives, (2) are longitudinal, prospective, and forward-looking, (3) are cross-nationally comparative, and (4) combine information on institutional context and individual behaviour. If we want to prepare Europe for the future ahead, we need data that are based on a life course perspective, which views the course people’s lives take as shaped by earlier life circumstances, the families and social networks in which people are embedded, and the institutional, cultural and economic context.
A breakthrough towards better understanding of the underlying forces behind the demographic developments can be achieved by integrated analyses that address both the societal (macro) level processes and the behavioural mechanisms on the micro level of individuals and households. Such analyses have to rely on appropriately designed surveys that collect micro-level data and on databases including the relevant macro-level data. Much of the necessary macro-level data is available and needs to be adequately integrated into the analysis programmes.
Pertinent differences in long-term demographic development, in the ways these societies are organised, in their cultural characteristics and in their family policies exist across Europe. Disentangling the causes of the differences in demographic reactions requires comparable data from many countries that represent a considerable variety of demographic, social, welfare, and cultural regimes. Such data could only be collected in consolidated international efforts with a high degree of standardization.
Longitudinal panel studies that consider life course experiences from the individuals’ past and relate them to the behaviours that follow can achieve crucial improvements in our understanding of demographic choices and their consequences. To make causal inferences, the analyst needs data where the hypothesized cause is observed before the outcome in a person’s lifetime. The panel design provides a possibility to use the broad range of information pertaining to the time of data collection in explaining subsequent behaviour that is captured in a following panel wave.
The Generations and Gender Programme was initiated in 2000 in response to the data needs identified above. Key features of the GGP initiative are:
• Cross-national comparability. The comparative focus allows analyses of the ways in which policies, culture and economic circumstances influence dependencies between men and women and between the young and the old.
• A longitudinal design. The GGP survey applies a panel design – collecting information on the same persons at three-year intervals – to allow the examination of causes and consequences of inequalities between genders and generations.
• A large sample size. The GGP survey has an average of 9,000 respondents per country, making it possible to study numerical minorities and uncommon events.
• A broad age range. The GGP survey has a focus on the whole adult population and thus includes respondents between the ages of 18 and 80, making it possible to address research and policy issues across the entire life course.
• The combination of micro and macro data. The Contextual Database should include country- and regional level data on demographic, economic, cultural and policy indicators.
• A theory-driven and multidisciplinary questionnaire. The development of the survey is inspired by a theoretical framework that emphasizes that key life-course decisions are influenced by attitudes, social norms and perceived and actual economic and institutional constraints.
Project Objectives
The main objective of the Design Study was to develop a European research infrastructure to ensure a sustainable knowledge base for population related policy formulation. To achieve this main objective, a number of more specific technical and scientific objectives had to be fulfilled. Below, these objectives are outlined.
Providing easy and well documented access to data. Data accessibility is of essential importance for maximising the impact of a Research Infrastructure. The objective of the GGP Design Study was to establish a system that informs the research and policymaking communities of the existence and potential use of the Infrastructure and provides easy access to both the macro and micro level data for the purpose of research. Providing information involves active presentations at appropriate gatherings of researchers and policymakers and inclusion of the relevant information in periodicals and other publications. A user-friendly main web entry point should be designed for a variety of users of this infrastructure, linking to documents of the underlying rationale, on specific technical information as well as to the data itself. Respecting appropriate measures for protecting personal data, appropriate screening and security procedures had to be developed for data access. In this context, the challenge was to ensure unobtrusive and quick data access, which is known to increase the research output based on a data made available by the Research Infrastructure.
Development of new measurement tools. The objective of the GGP Design Study was to improve the validity and reliability of existing measurement instruments and work out ways of collecting the necessary data more efficiently. In many topics envisaged as part of this Research Infrastructure, elaborate specialised measurement tools existed. The objective was to ensure that information allowing highly relevant and innovative analyses combining a number of issues pertinent to demographic behaviour could be collected in the framework of one survey at an acceptable level of respondent burden. Continuous development of measurement instruments contributes to higher data quality as well as to the completeness of gathered information on a specific topic. The measurement instruments should therefore aim towards constant improvement and to utilize new substantive and technical knowledge of the disciplines they are based in. The Design Study aimed at achieving a step-change in the measurement of economic well-being and social support networks as well as decision-making processes in the life course in the context of a large-scale multi-disciplinary survey.
Integrate new research areas. The GGP aims at including all topics with well-established theoretical relevance for explaining demographic behaviours and with a link to public policies. While this greatly facilitated developing survey modules based on a clear theoretical framework, several emerging areas of scientific inquiry into demographic developments had not been covered. Therefore, the Design Study had the objective of broadening the scope of disciplinary perspectives included in the GGP. This would involve developing measurement tools that would allow analysing the interplay of psychological traits, genetic and other biomarkers with demographic behaviours. Emerging research in these areas suggested a potential to contribute greatly towards our understanding of determinants of demographic behaviours and the Design Study integrated leading centres in these research areas, which would be in a position to develop an internationally applicable measurement standard.
Improvement in the quality of measurement instruments and the quality of survey data. In today’s context of decreasing rates of survey participation, securing high quality survey data is becoming more challenging. The objective of the GGP Design Study was to improve the existing GGP survey design to minimise distortions arising at different stages of the data collection process. This should be achieved through research in sampling methods, in methods of improving survey response and in developing data adjustment procedures, applied to the internationally comparable survey with longitudinal design. Continuous development of measurement instruments was expected to contribute to higher data quality as well as to the completeness of gathered information on a specific topic. The measurement instruments should aim towards constant improvement and utilize new substantive and technical knowledge of the disciplines they are based in.
Improvement in the cost-effectiveness of survey data collection. Declining rates of survey participation, increasing cost of personal interviewing as well as the emergence of new technologies drive the need to investigate alternative approaches to data collection. The objective of the Design Study was to improve the quality of the collected data while reducing response burden. Research activities in the field of data collection methods were expected to utilize technological advances such as Computer Assisted Interviewing (CAI) and the opportunities offered by the continuously growing access to the internet. It should furthermore investigate the use of information from administrative registers at different stages such as sample preparation, and data quality control as well as a supplemental source of data. Based on the results, recommendations should be made for the use alternative data collection methods. New findings should also contribute to better understanding of the impact of different modes of data collection and the use of the new technologies on data quality and the results were thus expected to contribute to the survey methodology as a scientific discipline.
Improvement of data comparability across countries. Comparability of data collected across different social, political and economic contexts is a fundamental building block of any cross-cultural study. Only coherent data based on common definitions, classifications and methodological standards can provide sound foundation for comparing different contexts and thus allow making inferences about the size and nature of the contextual effects to inform policy. The objective of the Design Study was to elaborate and implement measures to improve comparability of the internationally collected micro and macro-data. To achieve this, the implementation of the existing data collection instruments in different contexts was to be scrutinized and the collected information analysed.
Development of new methodological approaches in data analysis. The complexity of the integrated design of combining different sources of data as well as different disciplines and measurement instruments is inevitably translated into a complexity of the underlying data structures. The longitudinal nature of the survey design introduces additional considerations such as dealing with missing and incomplete information due to panel attrition. Complex, multilevel data structures require alternative approaches to data analysis. The objective of the Design Study was to investigate the possible approaches towards conceptualizing and testing innovative approaches in analysing incomplete complex data in a longitudinal perspective.
Advance understanding of the survey process. The GGP is the only large-scale multidisciplinary and internationally comparable panel survey, that encompasses the whole key family-building, working and early retirement age range (18-79). As such it presents an immense asset not only for its contributions to the understanding of contemporary demographic challenges, in particular for Europe, but also to possible advancements in survey methods. As a Research Infrastructure, the Design Study aimed to build on methodological excellence of its implementation as well as to contribute new knowledge to the field of survey methodology. The objective of the Design Study is to gain better understanding of the survey process, especially through an inter-cultural perspective, in order to reduce respondent burden and improve the quality of survey data.
Project Results:
In this section, the main scientific results and outcomes will be presented. It is organized around four topics, that reflect both the objectives of the GGP Design Study outlined in the previous section and the organization of the Design Study in a number of Work Packages. These four topics are:
1. Data Availability and Access
2. Quality of Research Instruments
3. Development of New Questionnaire
4. Development of New Data Collection Methods
Data Availability and Access
A key objective of the GGP Design Study was to provide easy and well-documented access to data. Within the Design Study, activities related to data access and documentation were concentrated in WP2 (Data Harmonization), WP3 (Data Repository and Dissemination) and WP4 (Maintenance and Development of Contextual Database). These three WPs were coordinated by the Working Group on Data Preparation and Dissemination.
The GGP Research Infrastructure provides access to micro-level data (collected through the Generations and Gender Survey) and macro-level data. Two types of access to micro-level data are available. The most extensive use – preferred by researchers – is through the analysis of f ully harmonized individual-level datasets that can be downloaded from a secure server by registered and approved users. A more descriptive way of assessing the data – preferred by journalist, policy makers, and researchers orienting themselves about the possibilities of the data – is by analysing the data on-line using the Nesstar analysis tool. Simple registration suffices to use this tool. Access to the macro-level data is through the on-line Contextual Database. In this section, we will describe results and developments concerning:
• Generations and Gender Survey
• Contextual Database
• Data Documentation and Access
Generations and Gender Survey
Micro-level data on intergenerational and gender relationships are collected through the Generations and Gender Survey (GGS). In this Section, we briefly describe the key features of the GGS, the activities undertaken to evaluate the quality of the data collected through the GGS, and the features of the next generation of GGP surveys, titled GGS 2015 and Beyond.
The GGS is a panel survey that focuses on the whole adult, non-institutionalized population, thus allowing it to address research and policy issues across the entire life course. Its longitudinal design, with information on the same persons being collected with a three-year interval, allows the examination of causes and consequences of the transitions individuals experience and the decisions they make. The average number of respondents per wave and per country is 9,000, allowing the GGP to study numerical minorities and uncommon events. Preferably, respondents are interviewed face-to-face, in order to ensure the highest possible level of cooperation. The GGP questionnaire is multidisciplinary and covers a broad array of topics of relevance to demographers, economists, sociologists, social psychologists and epidemiologists. At the same time, the development of the survey has been inspired by a unifying theoretical framework in which key life-course decisions are thought to be influenced by attitudes, social norms, and perceived and actual economic, social and institutional constraints. Key topics covered by the GGS include:
• Fertility
• Partnership
• Transition to adulthood
• Work-family balance
• Gender relations
• Life course and decision-making
• Intergenerational exchanges
• Informal and formal care
• Wellbeing and health
• Grandparenthood
• Economic activity
• Retirement
At the start of the Design Study, GGS data availability was still very limited. For a few countries, non-harmonized wave 1 datasets were available. Within the Design Study, harmonization procedures to ensure the quality and comparability of country-level data were established and refined, first for wave 1 data sets, and later on for wave 2 data sets. To ensure the cross-national comparability of the GGS, transparent and centrally coordinated harmonization activities are essential.
To achieve the highest level of cross-national comparability, the GGP combines both “input harmonization” and “output harmonization”. “Input harmonization” requires country teams that conduct the GGS to implement a centrally developed survey design and model questionnaire. As the details of the survey collection process are largely determined at the country level, and reflect national research interests and financial constraints, the national implementation of the GGS might deviate from the centrally suggested guidelines. These national-level deviations from the central guidelines and the model questionnaire, in combination with the need for rigorous quality control, require data harmonization procedures to be carried out in the post-processing phase. Such “output harmonization” is applied when the national implementation of the survey differs with regard to what is measured and how it is measured.
The initial post-processing activities of data collected by national survey organizations implementing the GGS are conducted by the national GGP teams. Subsequently, in order to assure maximal consistency and the highest possible level of standardization, the final part of the harmonization procedure is carried out centrally by the data harmonization unit of the Netherlands Interdisciplinary Demographic Institute (NIDI). This final part of the harmonization procedure is highly automated using advanced programming techniques in Stata®, a powerful statistical software package with great data manipulation capabilities. These harmonization activities not only check whether countries have consistently implemented the guidelines for pre-harmonization (i.e. variable naming and coding, country specificities, missing data), but also aim to increase the international comparability as well as the user friendliness of GGS data for the scientific community. These procedures include an extensive checking procedure of the routing of the questionnaire and the consistent coding of different types of missing data. Moreover, it encompasses the sorting of grids and event histories, and the construction of variables that contain basic socio-demographic characteristics of the respondent. The harmonization activities conducted at NIDI after the national GGP teams have submitted the pre-harmonized data files are described in detail in Deliverable 30 (Documentation of Harmonization Procedures).
Harmonization is a dynamic process. Since the release of the first harmonized data files in 2009, the scientific community of data users has been keen to communicate its feedback on these files via electronic mail or in personal communication during events such as the GGP User Conferences and other meetings. Moreover, Work Package 2 systematically analyzed the data quality of the GGS data and revealed ways in which the data harmonization procedures could be improved, for instance, in terms of routing checking procedures. Based on the feedback of users and these emerging new requirements, the user-friendliness of GGS data has been increased by evaluating and improving the harmonization procedures. As a result, two major revisions of Wave 1 data have been released during the term of the project. The first revision in October 2010 —Version 3.0—, included a major evaluation and improvement of the routing check procedure that had been moved from an XML environment to a Stata® environment. The second revision in June 2012 —Version 4.0— included a large number of additional constructed variables entailing socio-demographic characteristics of the respondent.
By the end of the Design Study, first-wave data have been made available in a harmonized format for sixteen countries: Austria, Australia, Belgium, Bulgaria, Estonia, France, Germany, Hungary, Italy, Lithuania, the Netherlands, Poland, Romania, Georgia, Norway and Russia. In addition, a dataset of the Turkish sub-sample of the German GGS is available. In 2012, the first second-wave harmonized datasets were released for six countries as well (Bulgaria, France, Germany, Georgia, Hungary, and the Netherlands). Thus, micro-level data on more than 157,000 individual respondents has already being made available.
Contextual Database
From the outset, the GGP has acknowledged that intergenerational and gender relationships are influenced by the demographic, economic, cultural and policy contexts in which individuals are embedded. Thus, a Research Infrastructure that provides scientists, policy analysts and journalists with data to study these issues, should provide both micro- and macro-level data. In line with this fundamental insight, the GGP has developed a Generations and Gender Contextual Database (CDB), next to the national Generations and Gender Surveys (GGS). The CDB serves two main purposes. First, it aims to support research on macro-level influences on micro-level processes, taking both individual-level and contextual-level characteristics into account. Second, it allows researchers to study macro-trends across time and space.
The data for the CDB are derived from two main sources. Where possible, existing online-databases by international organisations and research consortia are used. In addition, in each country participating in the GGP, a group of national experts has been set up to collect very specific indicators at the national and sub-national regional level, which are poorly represented in international databases (e.g. in terms of time coverage). The CDB coordination team carries out the harmonisation process of these data. In this process, each indicator derived from different national and international sources is cross-checked in order to select the best combination of data to be included. The choice of the data sources is determined by the following criteria: (a) compliance with CDB guidelines and international standards, (b) comparability across countries, (c) completeness, spatial and temporal availability of the respective indicator, and (d) availability of well documented metadata information and variable definitions. The process aims at maximising the geographical and temporal coverage of harmonised indicators for which data are made available.
In total, the CDB aims to include more than 200 cross-country comparative indicators, spanning a broad array of 16 different domains. The database includes quantitative indicators which in many cases cover not only national-level data, but also data on the sub-national, regional level. In addition, there are descriptions of main policy changes and reforms over time. In order to comply with the retrospective dimension of the GGS, most indicators cover time-series from 1970 onward or even longer periods of time. At the current stage, contextual data for around 100 indicators have been harmonised and made available in the GGP Contextual Database, covering up to 60 countries in Europe, North America, Asia and Oceania. The major areas covered by these indicators are: Demography, Economy and Social Aspects, Labour, Employment and Pensions, Child Care, Education, Health, Culture, and Tax and Benefits. These indicators can be accessed from the GGP website (http://www.ggp-i.org/) using a user-friendly interface that allows both on-line analysis of trends and easy downloading of indicators in multiple formats.
Overall, the GGP Contextual Database comprises a unique combination of features that might also serve as a blueprint for contextual databases of other surveys. This combination of features includes (1) harmonised cross-country comparative indicators for up to 60 countries combined from different sources, often covering not only the national but also the sub-national regional level; (2) extensive temporal coverage in order to support longitudinal analyses over time; (3) information on policy developments over time; (4) a dynamic user-friendly web environment, which has innovative functionalities both in terms of metadata documentation and of dynamic geo-coding of national as well as regional data in the extraction process. The geo-coding-function aims to support users to link the contextual data to the individual-level data for multi-level analysis. Different geo-code-systems are supported. This includes the system used in the GGS to identify place of residence of an interviewed person, as well as other systems such as NUTS- or OECD-codes. The coexistence of all these features in the GGP-Contextual Database makes it a unique support tool for researchers interested in the micro-macro linkages of social structures and processes.
At the end of the Design Study, the focus of activities is on increasing the number of harmonised policy indicators in the CDB. In this endeavour, close contacts have been established with other database initiatives working on similar aspects. This includes the Comparative Family Policy Database, whose indicators have recently been included in the CDB, the Population and Policy Database of Population Europe and the Multilinks Database on Intergenerational Policy Indicators. Another plan for the future development of the CDB is to include contextual information on changing social norms and attitudes at the national and regional level over time by aggregating response data from surveys such as the European Social Survey and the European Value Survey. In addition, opportunities for cooperation with coordination teams of other surveys (e.g. SHARE, ESS), which also recently started setting up contextual databases, will be explored. The GGP is very open to share its longstanding experience in the development of its GGP Contextual Database with other Research Infrastructures and to cooperate in the collection and dissemination of contextual data, as there is a substantial overlap in the contextual indicators provided by the GGP, ESS and SHARE. We believe that such cooperation will contribute to making the development of these databases more cost efficient and effective.
Data Documentation and Access
Providing easy and well-documented data access is essential to enable the widest possible range of users to benefit from the rich potential of a Research Infrastructure. Therefore, GGP users need quick access to information on the content of the data that the GGP provides, as well as its geographical and temporal coverage. Information on data quality and cross-country comparability has to be easily accessible as well. Quick and high-quality data access can maximise the impact of the GGP both on the research and on the policymaking communities.
GGP survey data and contextual data are made available free of charge through a single online data repository, which can be accessed through the GGP website (www.ggp-i.org). Several data access-options tailored to different audiences are available. Researchers can apply for access to harmonised GGS-data and download these data as soon as their application has been approved and access has been granted. A broader audience is targeted by the Nesstar online system , which allows a wide range of potential users, such as journalists, policy analysts and others, to analyse the survey data on-line by making tabulations and simple analyses. Moreover, the Contextual Database aims to serve both researchers and broader audiences by providing access to macro-level information on demographic, economic and political conditions.
Data access procedures are set up in compliance with Directive 95/46/EC of the European Parliament and of the European Council of 24 October 1995 on the protection of individuals with regard to the processing of personal data and on the free movement of such data. There are three levels of data access. First, full access to harmonized GGS micro-data files is granted to bona fide researchers who register to the online archive for any academic or policy-relevant research question. Second, just by providing a valid e-mail address, any user can gain access to the Nesstar system. Third, the GGP Contextual Database can be accessed without any type of registration.
Countries participating in the GGP provide de-personalised, clean and pre-harmonised individual-level data along with documentation in English for all survey instruments. They also enter into a legally binding agreement with the GGP which allows dissemination of the national surveys by the GGP Data Archive. Data release agreements have been concluded with all countries submitting data. Agreements regulate relevant data protection issues and outline the requirements for potential users who wish to apply for data use.
GGP aims at providing unobtrusive access to micro-data files while at the same time respecting appropriate screening and security procedures for protecting survey data. Since the amendment of the procedure in 2011/12, GGS data from most countries can now be accessed through a new, lighter application procedure where applicants have to sign just one form on “Statement of affiliation, confidentiality and acceptable use for GGS data sets” outlining the obligations to data users. Harmonised micro-data files are released to successful applicants electronically in SPSS and STATA formats. Users are alerted by e-mail about new versions of GGS data files. At the end of 2012, more than 500 projects have been or are using GGP data. The GGP bibliographic database contains information on more than 680 scientific publications that have been produced on the basis of GGS data.
The GGP data archive also provides users with harmonized background documentation on data. Good data documentation is paramount to effective data use. Researchers need metadata or “data about the data” to understand the numbers and transform them into meaningful knowledge. Metadata is essential to evaluate the cross-country comparability of data.
The metadata documentation in Nesstar provides GGS-data users with transparent and easy digestible user-friendly information on data collection and processing procedures. Metadata is provided in compliance with the Data Documentation Initiative (DDI) metadata standard . Through the GGP webpage, users can access a range of relevant metadata on the micro-level datasets, including:
• information about bibliographic citation, study scope and data access;
• information about national data collectors (i.e. producers, distributors, depositors, and funding agencies);
• information on data collection methodology and processing (i.e. sampling procedure, mode of data collection, type of research instrument, characteristics of data collection situation, actions to minimize losses, control operations, weighting, cleaning operations, and response rate);
• overviews of the cross-country availability of variables and descriptions of country-specific deviances from the coding scheme or the target sample;
• information on data entry and data pre-harmonisation by national data collectors, as well as information on central data harmonisation procedures.
Extensive metadata documentation on indicators in the CDB is available as well. Each indicator is linked to a “Meta information sheet”, which includes indicator-level information, such as definition, unit of measurement, time coverage, national data providers, sources used, and the geographical reference areas. In addition, specific metadata on each single data entry is offered. This includes the data source and its web-link, the retrieval date, information on any calculation or deviation from variable definition as well as on any potential break in series (e.g. due to revisions of data collections methods and/ or changes in country boundaries). This ensures a high level of transparency with regard to the origin and quality of the data.
With the implementation of new – simplified – procedures for data access in 2011, data access has been sped up. In the future, the GGP will continue streamlining data access procedures with the goal of releasing downloadable micro-level data to applicants as expediently as possible. The GGP Data Archive will pursue its efforts towards high quality and well-documented data, in compliance with the norms and standards promoted by the archives of the Council of European Social Science Data Archives (CESSDA). This is essential to increase the number of users of GGS micro-data files. In addition, the Nesstar online archive system and the Contextual Database will be maintained and extended, as soon as new data become available.
Quality of Research Instruments
Two key objectives of the GGP Design Study were to improve the quality of measurement instruments in general and the quality of data comparability across countries in particular. Within the Design Study, activities related to these objectives were concentrated in WP6 (Research on Methods of Measurement and Analysis). In this WP, the emphasis was on the reliability and cross-national equivalence of item scales. This WP was coordinated by the Working Group on Methodological Research. In addition, evaluation of questions on specific topics were evaluated within some of the WPs related to Measurement, in particular WP7 (Measuring Economic Well Being), WP8 (Measuring Social Support Networks), and WP9 (Life Course and Decision-Making).
At first sight, the main data quality issues in comparative research do not seem to differ much from those in one-country study. First, there are issues related to quality of the sample that is realized. Is the realized sample a fair representation of the underlying population it is thought to represent? If not, how large is the bias and can we develop methods to correct for the potential biases in the realized sample? Second, there are issues related to the quality of the data collected from the realized sample. Have all the data been collected or are parts missing for a large proportion of respondents? Are the data that have been collected reliable and valid indicators of the concepts that they are expected to measure? However, in comparative research, additional issues arise. It is expected that the data collected in the countries that are incorporated in the data set are comparable in such a way that if differences between countries in an indicator are observed, these differences can be attributed to country-differences in the substantive construct this indicator is expected to represent, rather than to differences between countries in the realized sample or in the reliability or validity of the indicators themselves. And very often, it is unclear whether this expectation is met, as there can be large differences across countries in sampling frames, sampling methods, interviewer modes, interviewer instructions and behaviour, translations of the questionnaire items, etc. This issue of comparative survey quality is of major importance to the GGP, as it is a key research infrastructure for comparative research. But to what extent does the GGS live up to its promises? The data quality of the GGS has been examined with a focus on a variety of aspects of comparative data quality.
With regard to sampling, the use of a probability sample was prescribed by the sampling guidelines, and it turned out that all countries complied to that prescription. The exact methods in which this was done, though, were country-specific. In most countries, a simple random sample was drawn from national registers or Census information, but in countries where this was not feasible, other randomized methods were employed.
Guidelines for GGP fieldwork implementation were not very elaborate, and as a result, most countries adhered to country-specific best practices. Two issues warrant specific attention from GGP users, though. First, the population covered by the survey (in particular the in- or exclusion of the institutional population , and the age range of the sample) differs somewhat between countries. It is suggested that users drop the small numbers of respondents that were outside the defined GGP universe of 18 to 79 years of age or were living in institutions. Second, the timing and duration of the wave 1 fieldwork differed substantially across countries. As a result, country differences could partially reflect period differences that operate across countries. Users should take this into account in interpreting differences between countries.
Response rates in GGP countries varied considerable, from a rather low 42% in Belgium to a very high 97% in Romania. On average, the response rate was about 67%. To put this response rate into perspective, it can be compared to averaged country response rates in other cross-national comparative surveys. The Survey of Health, Aging and Retirement in Europe (SHARE) reports an overall household level response rate for ten countries in wave 1 of 55% . In wave 3 of the European Social Survey (ESS), conducted in 2006, the averaged national response rate for 25 participating countries was 64%. Finally, the 2009 wave of the European Union Survey of Income and Living Conditions (EU SILC) reports a response rate for new households of 73% for 29 participating countries . Although it is hard to draw definite conclusions, given differences in sampling frames, survey mode, and survey agencies (e.g. official statistical offices versus private fieldwork agencies), these figures suggest that the response rate of the GGS compares quite positively with those of other major comparative surveys in Europe.
Even more important than the level of nonresponse, is the bias in the nonresponse. To evaluate that bias, survey data were compared to population data on a range of characteristics. In addition, this comparison was made both for the unweighted sample and for the weighted sample. For unweighted data, considerable bias was detected for such characteristics as age, gender, region, marital status, household size, and level of education. When the data were weighted using either the weights developed by the country teams or the centrally-developed weights, biases with regard to age, gender, region, and household size were substantially reduced. However, biases with regard to marital status and level of education remained. Therefore, users are advised to apply the weights provided in the GGS datasets whenever feasible. At the same time, caution is needed when the data are used to describe marital status and educational distributions of the population at large.
What implications should these findings have for future waves of the GGP? A first implication is to strengthen efforts to increase cross-national comparability. These efforts could include (a) alignment of the timing of waves in the participating countries, (b) more elaborate fieldwork guidelines, and (c) more elaborate and uniform reporting of contact and nonresponse information. Another important implication is to enhance efforts to document the representativeness of the realized samples and to improve uniformity in the weighting procedures. Such efforts could profit from better documentation of national weighting procedures and/or implementation of central weighting procedures.
An important issue in comparative research is the cross-national equivalence of multiple-item sets. Multiple-item sets are mainly used to measure individual scores on theoretical constructs that are not easy to measure, like human values or ethnic stereotyping. Asking multiple questions on the same theoretical construct – a latent variable in the language of Structural Equation Modelling – allows one to test whether individuals answer the same set of questions in the same way. Only if they do so to a considerable extent is it warranted to compare scores across individuals, or – in the case of cross-national equivalence – across countries. In the Design Study the cross-national equivalence of all multiple-items scales included in the GGS has been examined, and it is concluded that only one of them – the De Jong Gierveld loneliness scale – qualifies for the highest level of equivalence – called scalar equivalence – across almost all countries. Many other scales exhibit partial equivalence, which implies that scale scores are not completely comparable across countries. This may sound as ‘bad news’, but (a) the scales in the GGS actually perform quite well compared to scales in other comparative surveys, and (b) often a remedy is available by analyzing GGS data using Structural Equation Modelling (SEM). New developments in SEM now allow multi-country and multi-level analyses, as well as the application of SEM in hazard analyses, increasing its potential for demographers.
Another aspect of data quality is the extent to which all respondents who have to answer a specific question, actually do so. This is usually referred to as the issue of item nonresponse. Item nonresponse can result from the unwillingness of the respondent to provide an answer, for instance to a question he or she experiences as being too private, or from decisions of the interviewer not to pose a required question. In a comparative survey like the GGS, it could also be that country teams that are responsible for the implementation of the GGS decide not to include specific questions or to pose them to a particular selection of respondents only, in that way deviating from the guidelines provided in the GGS model questionnaire. On average, item non response in wave 1 GGS data sets was about 10 per cent. A detailed analysis of specific sub-sections of the GGS questionnaire suggested that decisions by country-teams to pose certain questions to smaller subsets of respondents than proposed by the GGS model questionnaire was a major factor in explaining the high item nonresponse in some GGS sections. In addition, the role played by question characteristics in explaining the item nonresponse patterns has been examined. Most question characteristics did not have strong effects on item nonresponse. Exceptions were questions on other persons (so called proxy questions) and questions about future expectations, that showed a somewhat higher level of item nonresponse.
Special attention has also been paid to assess the quality of specific sets of indicators. One such important set are variables that can be used to measure individuals’ and households’ level of economic wellbeing. It is important to bear in mind that unlike the EU-SILC, the GGS is not designed to provide extensive information about economic wellbeing. Whereas the EU-SILC is the source of information for assessing living conditions in the European Union, the key focus of the GGS is on gender and intergenerational relationships. Still, it has been assessed to what extent – given limited measures –GGS variables reflecting economic wellbeing resemble those of the EU-SILC and hence their usefulness in terms of measuring economic wellbeing for different demographic constellations and different age groups. A key finding is that the GGS with its focus on demographic trajectories and relations between genders and generations offers an important contribution towards assessing the life-course and economic outcomes. Moreover, the GGS is important in the sense that it has a longitudinal design. With consistent measures of economic wellbeing, we are not only able to assess how trajectories may have an impact on current economic wellbeing, but we are also able to assess how demographic changes between waves are related to changes in economic wellbeing. This design will provide important insights that cannot be done with the EU-SILC.
Based on the household income as reported by the respondent, poverty rates were computed by taking a standard approach similar to that used by EUROSTAT using EU-SILC data. Our estimates show that in those countries where household income is given by exact amounts, the estimated poverty rates are very similar to those of EU-SILC. The exceptions are Germany and France, where the GGS poverty rates are somewhat higher. The analysis shows that this is not due to the imputation procedure implemented. Most likely it is a result of the way household income is reported in income bands. The other exception is Hungary where the number of missing values for household income is large. Here the imputation does impact the poverty rates, and our recommendation here is to stick with the original income measure if the aim is to produce reliable poverty statistics. In addition, analyses showed that measures of material deprivation are highly consistent with EU-SILC. Descriptive statistics show that levels of deprivations are consistent with the subjective measures and also estimated poverty rates in that they move in same direction when the levels varies for different household constellations. The measures are also consistent in terms across country levels. France and Germany are the countries with lowest levels of economic deprivation – not matter how it is measured, whereas it is highest for Georgia. Overall, the conclusion is that the economic wellbeing measures in the GGS are of decent quality, but that country differences need to be taken into account when comparative analysis is done. We also feel that the inclusion of these measures provide high value added compared to EU-SILC – especially because of the longitudinal design and its emphasis on demographic processes.
A key innovation of the GGS is its use of the Theory of Planned Behavior (TPB) to guide the selection of questions on intentions concerning major life-events like leaving home, union formation, parenthood and retirement. Questions were not only posed on people’s intentions regarding these behaviors, but also about their attitudes, norms and perceived behavioral control concerning these behaviors. Analysis of the measures included in wave 1 of the GGS showed that the TPB, which forms the basis of the decision-making item sets, is a good framework for continued study of life course decision-making, not only because it is a sound and well-tested theory, but also because it provides a unifying framework for the factors proposed in other literatures to influence life course decision making and it offers valuable guidelines for definition and measurement of cognitions that contribute to life course decisions. It has already been used with success in studies of fertility intentions and realisation of those intentions, particularly those studies conducted as part of the REPRO project. Analyses of the TPB batteries for the other decisions included in the GGS show that the TPB also provides a good explanation of intentions for these actions. Based on the extensive analyses of TPB items, a number of recommendations were made for the development of the new questionnaire. The key recommendation in that respect is the replacement of the entire battery of perceived behavioural control questions for all decisions, in order to directly address the concept of control as defined in the TPB.
The social network indicators in the GGS have also been extensively evaluated. It turned out that the role-relation approach used (i.e. questions ask whether specific types of network members, e.g. partner, children, parents, friends, co-workers, provide and receive specific types of support) worked well, but had one important drawback; it was not possible to link support exchanges to specific members of the network. For instance, if someone has three children, one knows whether these children provide emotional support or not, but not who of these children provide support and who do not. It is suggested to add questions on specific support-providers and receivers in the new GGS questionnaire.
Development of new questionnaire
Many of the activities described in the previous section, directly fed into the development of a new questionnaire for future waves of the Generations and Gender Survey. This new 2015 GGS Questionnaire is one of the key deliverables of the Design Study (D26). A lot of efforts have been put in the preparation of this questionnaire, including:
1. Evaluating the quality of measurement instruments in the GGS wave 1 questionnaire (see the previous section);
2. Assessing the usage and usefulness of different parts of the GGS wave 1 questionnaire;
3. Identifying ‘blind spots’ in our scientific and policy knowledge and how to operationalize these;
4. Testing new questions and different forms of answer formats;
5. Adapting the flow of the questionnaire.
The task of evaluating the previous questionnaire and developing a new questionnaire was a combined effort by, on the one hand, sets of researchers who evaluated the existing parts of the questionnaire from a substantive point of view, and a set of researchers that focussed on the methodological aspects of the questionnaire, on the other hand. Substantively, the development of the new questionnaire focussed on measures of economic well-being (WP7), measures of social support and intergenerational relationships (WP8), measures on key life course decisions, such as leaving home, union formation, parenthood, and retirement (WP9), and measures of psychological traits and social well-being (WP10). Methodologically, the focus was both one the quality of individual measurement instruments, and on the general issue of developing a questionnaire that could be used in a mixed-mode approach to data collection, and yields equivalent responses in face-to-face, Web and telephone modes.
The key decision about GGS 2015 was that the focus should be on continuity. The value of a panel design largely rests on the ability to examined change across personal and historical time, and this is only possible if the same measurement instruments are repeated in subsequent waves of the survey. Thus, measurement instruments were only changed if there were strong methodological reasons to do so. At the same time, some level of innovation is necessary as well, as new topics emerge as being important from either a scientific or policy point of view. Finally, the content of the questionnaire is limited by the length of the questionnaire. It was decided that the length of the GGS 2015 questionnaire should not exceed that of the previous GGS questionnaire. Therefore, space for innovation could only be created by reducing the length of other parts of the questionnaire.
Compared to the previous GGS questionnaire, an important decision was to drop the use of optional modules. In Wave 1, four optional modules (on ‘nationality and ethnicity’, ‘previous partners’, ‘intentions of breaking up’, and ‘housing’) were included in the GGS. However, these modules generally were fielded by only a small number of countries, thus reducing the potential for cross-national comparison.
The new questionnaire was developed within WG3 on questionnaire development. It met four times. The work started with a discussion of the evaluation of the quality of the measurement instruments in Wave 1. Based on this discussion, decisions were made on which parts of the questionnaire needed to be revamped, and which parts could basically be left unchanged. Next, proposals for new questions were developed by the responsible substantive sets of researchers. Based on their proposals, a preliminary version of the questionnaire was prepared and discussed in WG3. Cuts were made to the draft version based on substantive and methodological considerations, but also based on the necessity to limit the duration of the survey. This draft version was discussed with the GGP Consortium Board and the Network of National Focal Points. Next, the draft version was tested in a pilot study in Slovenia. Based on the results of the pilot, the questionnaire was once again discussed in WG3 and adjusted. The adjusted version was then approved in the Consortium Board. The resulting questionnaire is available as D26.
Substantially, the 2015 wave of the GGS has retained its focus on understanding intergenerational and gender relationships. This includes attention to leaving the parental home, the family formation process, the combination of employment and intergenerational care obligations – both towards parents and children, and retirement. At the same time, GGS 2015 contains some innovative aspects. These include:
• A revamping of the social network module. The new social network module makes it possible to more clearly distinguish the support relationships between genders and generations. This is important to better understand why certain family members offer and receive support, whereas others do not.
• Better measures of how societal constraints and social policies influence the decisions that individuals make concerning family formation and retirement. This is important, as one of the main goals of the GGP is to be able to understand how people react to changes in policy environments.
• Inclusion of a personality module. This is important as personality can have an impact on how individuals behave across a wide range of life domains and can have a long-lasting influence on how individuals react to adversity in the lives.
• Revamping of questions on perceived behavioural control regarding leaving home, union formation and parenthood. These questions are part of our Theory of Planned Behaviour approach towards understanding demographic decision-making. However, the questions in the first waves of the GGS did not work as expected, and have been reformulated in collaboration with Icek Ajzen (the designer of the TPB approach).
Development of mixed-mode approach
A central aspect of the design study has also been to evaluate the survey design of the GGS. In general, the survey design of the GGS is excellent, but relatively expensive. The mode of data collection in the first waves of the GGS – face-to-face interviewing – is the gold standard in the field, but is much more expensive than other modes such as telephone or Web. In addition, alternative modes of data collection – and Web-questionnaires in particular – are becoming more popular and allow a substantial reduction in costs. To assess the potential of alternatives to face-to-face, a pilot study was conducted in which alternative data collection systems were compared in terms of data quality and effectiveness. The pilot study was organized in Slovenia. Based on that study and prior considerations, it has been decided to opt for a balanced combination of different survey modes that combine cost-effectiveness and high quality.
The key outcome of that evaluation is the adoption of a mixed-mode survey design, with first-time respondents being surveyed by use of a face-to-face interview, and repeat respondents preferably being surveyed by a web-based questionnaire. The use of web and telephone modes leads to a reduction of overall survey costs. First-time respondents are surveyed via a face-to-face interview as this increases their retention rate. Special care is taken to allow for comparisons of results across different modes.
While a mixed-mode approach will not necessarily be more efficient or cost effective in the initial survey, both the pilot study experiment and previous research indicate that an online access panel strongly suggests a more favourable situation in subsequent waves of GGS. Here, an initial face-to-face contact enables a higher persuasion and legitimacy of the survey. Once this is established and a person accepts to participate in subsequent waves of the survey they will be significantly more willing to participate via a web portal. Moreover, higher response rates to web survey in subsequent waves will greatly reduce overall variable costs of data collection as lower number of face-to-face or telephone interviews will be required. Such reductions in costs can then free up resources for other aspects of the research and hence lower overall costs. Country-specific differences are, however, expected, meaning that cost-reduction levels could differ between countries.
For the longitudinal component of the data collection, it is recommended that the interval between subsequent waves remains three years. This strikes a balance between sample attrition, which would likely be lower with shorter intervals between waves, and the scientific issue of needing enough time for changes in social and demographic behaviour and experience to occur. This, however, requires considerable due diligence in panel maintenance to ensure that attrition is minimized. Here, we recommend the adoption of five practices: a) continuous and close cooperation between the research institute, the fieldwork agency, the interviewers and the respondents; b) incentives for respondents that are of meaningful monetary value and will vary across countries; c) regular contact with respondents through letters, information brochures, requests for updated contact information, and where feasible, the collection of annual information via a short-form questionnaire; d) where possible, interviewer continuity is recommended to help establish a relationship between the respondents and the interview staff; and e) specialized interviewer training and supervision is essential
The overall recommendation is therefore to conduct the first wave with face-to-face survey followed by the web option and starting with the web option in all subsequent waves at three year intervals. If necessary and cost feasible, face-to-face and telephone surveying can be used as alternative modes.
In addition to the general recommendation to move to a mixed-mode survey design within the GGP, a number of additional recommendations were made, including:
• Full computerization of the questionnaire. With computerization, interviewer’s burden decreases, there is less data cleaning, coding and editing in the post-fielding phase, and data quality increases. In addition, centralized computerization of the questionnaire increases control over the questionnaire design and survey implementation in individual countries and lowers costs.
• The GGS keeps its focus as starting from a random sample of the non-institutionalized population, aged 18 till 79, with an average sample size of at least 9,000 respondents per wave. This large sample is necessary because the sample covers the whole adult life span and allows not just for cross-national analysis, but also for analysis at the national and sub-national level.
• The panel character is maintained, but the panel is to be refreshed regularly, to ensure an average sample size of 9,000 per wave. In the refreshment sample, young age groups that were not represented in earlier waves of the GGS, will be oversampled. Weights will be calculated to ensure both longitudinal and cross-sectional representativeness of the resulting dataset.
• The three-year interval between waves is retained as well. This strikes a balance between sample attrition, which would be lower with shorter intervals between waves, and the scientific need for sufficient demographic and socio-economic changes to occur. This three-year interval requires considerable due diligence in panel maintenance to ensure that attrition is minimized. A series of good practices to minimize attrition are proposed.
• A feature of the GGS that is cause for concern is the fact that surveys are conducted at different times in different countries. This lowers the opportunities for cross-country comparison. To increase the temporal comparability, the questionnaire for future waves of the GGP will be regularly updated, and countries are required to field that questionnaire within a limited time-window. In addition, only those country datasets that include at least 70 per cent of the questions in the GGP Standard Questionnaire will be included in the GGP.
Potential Impact
Overall, the GGP is developed to achieve two major types of impact. The first impact is scientific. The combination of surveys and contextual databases provides solid scientific knowledge about how major societal changes shape social and economic inequalities between genders and generations. The key features of the GGP are functional in helping to achieve this impact, such as the inclusion of adults of all ages, the large sample size, the cross-national comparability and the longitudinal focus.
The second impact concerns policy, both at the national and the European level. The GGP is geared towards optimizing not only the quality of empirical knowledge, but also that of policy recommendations. Moreover, the GGP can fill the knowledge gap that exists with regard to the Central and East European countries. By identifying links between policy regimes and generational and gender inequalities in income, health, support, and well-being, the GGP extends the substantive understanding of the risks of social exclusion, social isolation, and poverty. In addition, by organizing meetings with stakeholders and end-users, the infrastructure contributes to the development of new policies that are required to compensate for the potentially negative effects of demographic changes.
Scientific Impact
In this section, we assess the progress made in terms of the generation of new data and the stimulation of new research. We start by describing the numerical increase in data sets, data users and research output. Next, we briefly comment on the content of the generated knowledge.
Exploitation & Data Access
GGP survey data and contextual data are made available free of charge through a single online data repository, which can be accessed through the GGP website (www.ggp-i.org). Several data access-options tailored to different audiences are available. Researchers can apply for access to harmonised GGS-data and download these data as soon as their application has been approved and access has been granted. A broader audience is targeted by the Nesstar online system , which allows a wide range of potential users, such as journalists, policy analysts and others, to analyse the survey data on-line by making tabulations and simple analyses. Moreover, the Contextual Database aims to serve both researchers and broader audiences by providing access to macro-level information on demographic, economic and political conditions.
The first 7 harmonized GGP wave 2 datasets are now available and a total of 16 harmonised wave 1 datasets, with a total of more than 157,000 respondents have been updated or newly released. New data collection efforts have been started in Sweden, while Austria, France, the Netherlands, Italy, Lithuania, Russia, Australia and Hungary have collected subsequent waves of data. GGP data are available to the international scientific community through a single online data repository, which can be accessed through the GGP website (www.ggp-i.org). Over the years, the number of research projects using GGP data has steadily increased, reaching more than 700 in early 2013 (see fig 1.). In addition, the Nesstar online system allows a wide range of potential users, such as journalists, policy analysts and others, to analyse the survey data on-line by making tabulations and simple analyses. The Contextual Database (CDB) that provides access to macro-level information on demographic, economic and political conditions has been expanded and developed. Currently, the CDB includes more than 100 quantitative indicators. In order to comply with the retrospective dimensions of the GGP, most indicators cover time-series from 1970 onward or even further back.
The data provided by the GGP Research Infrastructure are increasingly utilised by the research community. Figure 2 shows the increase in publications based on GGP data. By the end of 2012, more than 1,000 papers and reports based on GGP data had been published or presented at conferences . This includes more than 617 articles in peer-reviewed scientific journals and over 30 PhD theses. The number of conference presentations using the GGP has been growing at a very steep rate and has reached over 450 recorded appearances and is set to grow further when Milan hosts the Second User Conference of the GGP in October 2013. As these numbers reflect, the GGP’s research output is growing and will continue to do so as more longitudinal and comparative data become available.
Existing research based on GGP
GGP data are used by individual researchers studying a specific topic in a specific country, but also by large consortia of institutions and researchers studying a broad subject in a comparative perspective. Below, some of the major projects that have used or are currently using GGP data are briefly described:
• MULTILINKS (How demographic changes shape intergenerational solidarity, well-being, and social integration: A multilinks framework) is an EU-FP7 funded project in which nine scientific institutions collaborated. The objective of MULTILINKS was to investigate how changing social contexts, from macro-societal to micro-interpersonal, affect social integration, well-being and intergenerational solidarity across different European nations. Using GGP data, the project has provided a wealth of new knowledge on intergenerational co-residence, well-being of older adults, and cross-national differences in generational policies.
• REPRO (Reproductive decision-making in a macro-micro perspective) is another EU-FP7 funded project in which ten scientific institutions collaborated to generate new scientific and policy-oriented knowledge on the factors that drive changes in birth rates and influence reproductive decision-making of contemporary Europeans. GGP data were used to study how childbearing intentions and decisions of Europeans were influenced by attitudes, social norms and perceived and actual constraints.
• FamiliesAndSocieties (Changing families and sustainable societies: Policy contexts and diversity over the life course and across generations) is an EU-FP7 project in which 25 institutions collaborate. The project will explore the growing complexity of family configurations and life course transitions across and within European societies and examine their implications for the well-being of children, women and men, inequalities in life chances, intergenerational relations and care arrangements. The project will also investigate how policies address family diversity and its consequences. The aim is to assess the compatibility of existing policies with the increasingly diverse and complex family life courses in Europe, and to contribute to evidence-based policy-making.
• CONOPP (Contexts of Opportunities. Explaining cross-national variation in the links between childhood disadvantage, young adult demographic behaviour and later-life outcomes) is an ERC Advanced Grant project led by Aart Liefbroer at the Netherlands Interdisciplinary Demographic Institute in which GGP data are used to examine the role played by demographic events and trajectories in producing and reproducing inequalities. It studies cross-national variation in the strength of the relationships between childhood social disadvantage, demographic decision-making during young adulthood and subsequent economic, social and health outcomes. The project relies heavily on the use of data from the GGP Research Infrastructure.
• Families in context (Families in context: Unravelling the ways in which policy, economic, and cultural contexts structure generational interdependencies in families and their life outcomes) is an ERC Advanced Grant project led by Pearl Dykstra at Erasmus University Rotterdam. The project’s main objective is to unravel the ways in which policy, economic, and cultural contexts structure intergenerational dependencies in families. Adopting a multigenerational view of family ties, across life phases, it compares and contrasts different theoretical models for why members of different generations help each other.
• FAMHEALTH (Family life courses, intergenerational exchanges and health and well-being in later life), an ERC Advanced Grant project led by Emily Grundy at Cambridge University, will use GGP data, for example, in analyses aimed at uncovering whether country level factors moderate associations between early fertility and later life health and to investigate mid-life (age 35-55) adults’ provision of support to parents and associations with health.
In addition, several other EU-FP6 and FP7 projects, such as MAGGIE, MICMAC and NEUJOBS, as well as a number of projects by ERC Starting (e.g. Arnstein Aassve, Helga de Valk, Letizia Mencarini, Jan van Bavel, Brienna Perelli-Harris) and ERC Advanced Grants (e.g. Gøsta Esping-Andersen) are using or have used data from the GGP.
Policy Impact and Dissemination
A major aim of the GGP is to disseminate findings based on its data to a wider audience of policy makers as well as to the general public. Since 2012, GGP at a Glance, our bimonthly newsletter is distributed among over 500 subscribers. Each Newsletter includes an item called Did you know? that alerts the reader to recent scientific and policy-relevant GGP findings. In addition, abstracts are published from articles based on GGP data that have appeared in peer-reviewed journals. Finally, GGP-related news is included in the Newsletter. In addition to the Newsletter, the recently launched GGP Research Notes series contributes to the broader discourse on contemporary population challenges and demonstrates what GGP data can contribute to policy and research agendas. Three to four issues per year are published, in which internationally renowned scholars discuss GGP research with clear policy implications.
In order to further develop the GGP Community, GGP User Conferences were held in Budapest in 2011 and in Milan in 2013. This 2 day conference saw the presentation of over 50 pieces of original research using GGP data. A 3rd User conference is planned for 2015 and the GGP will encourage both the use of GGP data and the existing findings of GGP research through international conferences and policy networks. As part of these efforts, in 2014 the GGP coordination team will participate in the Berlin Demographic Forum and European Population Conference, providing a bridge between scientific research and policy makers.
In the future, we aim to organise policy events – both small scale ones, like targeted seminars, and large scale ones, like policy conferences –to disseminate policy-relevant knowledge that is generated by the GGP. The GGP will also work with the United Nations Economic Commission for Europe (UNECE). The UNECE views the GGP as a means to inform policy making about ageing and gender issues. In 2008, a special conference on ‘How Generations and Gender Shape Demographic Change’ was organized by the UNECE to discuss first results based on data from the GGP. Official representatives of 32 UN Member States participated. Dissemination activities will also benefit from the involvement of Population Europe (PE), the collaborative network of leading demographic research institutes and over 200 experts. With support of the European Commission, Population Europe has developed tools to disseminate research outcomes to policy-makers, civil society actors, the media, and lay audiences. Involvement of Population Europe will also enable the use of PE Digest, Population Europe’s series of research briefs that summarize policy-relevant results from scientific research published in international journals. Finally, our aim is to expand our website with GGP material of interest to policy makers.
Providing easy and well-documented data access is essential to enable the widest possible range of users to benefit from the rich potential of a Research Infrastructure, including policy makers and other stakeholders. Therefore, the GGP needs to provide quick access to information on the content of the data that the GGP provides, as well as its geographical and temporal coverage. Information on data quality and cross-country comparability has to be easily accessible as well. Quick and high-quality data access can maximise the impact of the GGP both on the research and on the policymaking communities. With our online analysis suite, the GGP has taken the first steps towards this within this design study.
The GGP provides a major research infrastructure that informs policy debates and decisions for the EU and its member states. Among the critical debates requiring this input are intergenerational solidarity, ageing, the nature of family structures, newly emerging social inequalities, work life balance and low fertility; moreover such debates have to be examined in the context of the fiscal crisis, employment patterns, interpersonal support mechanisms, and existing policy and structural constraints. A properly supported, sustainable GGP is of crucial importance for both social science research and policy formulation for Europe and its institutions.
Potential Impact:
The main aim of the GGP Research Infrastructure has been and will remain to provide social science data that can improve our scientific understanding of key societal challenges and to disseminate that knowledge to the scientific public via books and articles in scientific journals and to policy makers and the general public through newsletters and presentations at policy-related meetings. Below, we will first outline the kind of challenges to which the GGP provides answers, and how the GGP relates to major international research agendas. Next, we will provide an overview of the main activities that have been conducted during the FP7 project. Finally, we will outline plans for the future.
Societal challenges
Although the current economic crisis has led to a sharp increase in unemployment, long-term trends still indicate the emergence of significant labour force shortages. One way of resolving labour market shortages is increasing the labour force participation of women. To do so, the relationship between work and family life must be considered and implies a fresh look at how women and men divide tasks in the employment and family domains. The GGP contains information on educational and work careers, division of labour within the household and on preferences of men and women about this division of labour, that allow us to explore the coherence of career choices and the impact of macro factors such as the financial crisis, but also micro factors such as relationship developments, the arrival of a new child or the emergence of care requirements of dependent relatives. For example, the GGP helps us understand the gender specific challenges of employment and career advancement of highly educated and skilled women such as care arrangements and career breaks.
Another important issue is to increase the labour market participation of older workers. The GGP covers the whole life span and enables specific analysis of older workers within the economy. Specific questions regarding retirement intentions are included within the questionnaire and the longitudinal nature of the dataset allows for comprehensive mapping of retirement plans. Detailed information on the nature of work such as the type of contract, supervisory responsibilities, flexible hours and capacity to work from home are all included in order to enable policy makers and researchers an opportunity to understand the strategies for managing care responsibilities and the transition to retirement.
Population ageing has led to discussions on how intergenerational solidarity between younger and older generations can be maintained. It also has led to discussions about increasing demands on adult children to care for their frail parents. The GGP allows us to examine issue of support exchange not just from the point of view of the older generation, but from that of the younger generation as well. It will also allow us to examine under which conditions support exchange between parents and children has positive consequences for the well-being of support provider and support recipient, and under which conditions these consequences are negative.
The process of globalization has led to an emphasis on the flexibilisation of the labour market. This has weakened the position of young adults, a process that is exacerbated by the current economic crisis. As a result, social inequalities in the Netherlands and across Europe might increase. What will be the short- and the long-term impact of the recent (and in some countries continuing) financial crisis on individuals? In particular, and considering that the youth unemployment rate in the Eurozone stands currently at 24%, what will be its impact on the transition to adulthood, that is, on young people’s ability and decision to form an independent household, form a stable partnership, and enter parenthood? And how will the lives of these young adults compare to those of their parents? And how will they differ for men and women? The GGP can provide answers to these questions.
The process of ageing, and the concurrent partial retreat of the welfare state, makes the issue of healthy ageing a pertinent one. The GGP starts from the assertion that understanding ageing issues asks for a focus on the whole life course (what are the precursors of health in old age?), needs attention to the social networks surrounding people (who cares for whom?), and requires a broad conceptualization of well-being (social, economic and medical aspects need to be included). For example, loneliness has long been a topic that has featured heavily within the research outputs of both the NKPS and the international GGP . For society as whole the value of this research is also in the identification of individuals lacking in such relationships and how they might be supported in times of need and support.
The GGP also enables high quality policy research and analysis that are essential for efficient and effective government. Firstly, the multilevel nature of the GGP, enables for the reconceptualization of the welfare state using comparative data at the macro and micro level. Using the Contextual Database which includes policy information for all GGP countries, it is possible to map constellations of welfare arrangements across 19 countries. This will allow for a historical understanding of the Netherlands positioning within past and future welfare state constellations and the ability to contextualize the micro level survey data within these developments. In addition to this the GGP is capable of facilitating evaluations across a broad range of social and economic policies. At a time of austerity and in a climate where efficiencies are a priority, the crucial question for government and civil society is which policies work and which don’t. The detailed policy information available from the registers provided by CBS will enable the GGP to be used in complex policy analysis. For example, it will be possible to follow care arrangements for infirm spouses as the policy mix for elderly care changes in nature. Employment, housing, health & care, social insurance and education are all policy areas that could benefit from analysis of GGP data as it provides an economic and scientifically sound alternative to the gold standard of experimental policy studies.
Links to key policy and research agenda’s
Europe 2020, the vision document published by the Commission, is the European Union’s growth strategy for the coming decade. In a changing world, the EU aims to become a smart, sustainable and inclusive economy based on high levels of employment, productivity and social cohesion. The vision document notes that a successful EU 2020 strategy must be built on a good analysis of the constraints facing policy makers and acknowledges the challenge presented by European demography. The changing demographic landscape, with a lower proportion of young people in the overall population, is expected to result in substantially reduced potential growth by 2020. The crisis has exacerbated the long-term social challenges that Europe faces today, such as solidarity between generations in the context of an ageing society, social exclusion and child poverty, and the integration of an increasing immigrant population. Europe 2020 emphasizes that new policies must demonstrably contribute to social cohesion, tackling unemployment and fostering social inclusion of young and old, and men and women. The EU 2020 strategy is also is an agenda for all Member States, large and small, old and new, highly developed and still developing. Based on Europe 2020, the Commission has developed its research framework programme for the period 2013-2020, called Horizon 2020. Some of the key societal challenges that Horizon 2020 distinguishes are ‘Health, demographic change and wellbeing’, ‘Europe in a changing world: inclusive, innovative, and reflective societies’, and ‘Secure societies’. With its wide coverage of European countries and focus on the relations between generations and genders, the GGP is ideally suited to address these social challenges.
Population ageing has been a major focus within the United Nations Economic Commission for Europe (UNECE). In the Madrid International Plan of Action on Ageing (2002), and subsequent the Ministerial Conferences on Ageing (2008 and 2012), several key commitments and action strategies have been formulated. Advancing gender equality and the empowerment of women became a goal of the United Nations shortly after the Organisation was established. The first meeting of the UN Commission on the Status of Women was held in 1956. More recently, the vital importance of a life-course approach and of cross-nationally comparable survey data was stressed in the UNECE’s Chair summary of the 2013 Regional Conference on population and development . In particular, a life-course approach was deemed necessary to better understand how people’s lives’ unfold as they move in and out of partnership, as they experience various family transitions, and as they are confronted by unequal gender norms. Thus, it should come as no surprise that the UN has been the main driving force in establishing the GGP. They view the International GGP as a means to inform policy making about ageing and gender issues (see the Letter of support by UNECE).
Main activities during the FP7 project and future outlook
Providing easy and well-documented data access is essential to enable the widest possible range of users to benefit from the rich potential of a Research Infrastructure. Therefore, GGP users need quick access to information on the content of the data that the GGP provides, as well as its geographical and temporal coverage. Information on data quality and cross-country comparability has to be easily accessible as well. Quick and high-quality data access can maximise the impact of the GGP both on the research and on the policymaking communities. This Section describes how the GGP assures this kind of easy data access, and outlines plans for the future.
Data access
At the end of 2012, more than 500 projects have been or are using GGP data. The GGP bibliographic database contains information on more than 680 scientific publications that have been produced on the basis of GGS data.
GGP survey data and contextual data are made available free of charge through a single online data repository, which can be accessed through the GGP website (www.ggp-i.org). Several data access-options tailored to different audiences are available. Researchers can apply for access to harmonised GGS-data and download these data as soon as their application has been approved and access has been granted. A broader audience is targeted by the Nesstar online system, which allows a wide range of potential users, such as journalists, policy analysts and others, to analyse the survey data on-line by making tabulations and simple analyses. Moreover, the Contextual Database aims to serve both researchers and broader audiences by providing access to macro-level information on demographic, economic and political conditions.
Main Nesstar interface: Example
Data access procedures are set up in compliance with Directive 95/46/EC of the European Parliament and of the European Council of 24 October 1995 on the protection of individuals with regard to the processing of personal data and on the free movement of such data. There are three levels of data access. First, full access to harmonized GGS micro-data files is granted to bona fide researchers who register to the online archive for any academic or policy-relevant research question. Second, just by providing a valid e-mail address, any user can gain access to the Nesstar system. Third, the GGP Contextual Database can be accessed without any type of registration.
Countries participating in the GGP provide de-personalised, clean and pre-harmonised individual-level data along with documentation in English for all survey instruments. They also enter into a legally binding agreement with the GGP which allows dissemination of the national surveys by the GGP Data Archive. Data release agreements have been concluded with all countries submitting data. Agreements regulate relevant data protection issues and outline the requirements for potential users who wish to apply for data use.
GGP aims at providing unobtrusive access to micro-data files while at the same time respecting appropriate screening and security procedures for protecting survey data. Since the amendment of the procedure in 2011/12, GGS data from most countries can now be accessed through a new, lighter application procedure where applicants have to sign just one form on “Statement of affiliation, confidentiality and acceptable use for GGS data sets” outlining the obligations to data users. Harmonised micro-data files are released to successful applicants electronically in SPSS and STATA formats. Users are alerted by e-mail about new versions of GGS data files.
Data documentation
The GGP data archive also provides users with harmonized background documentation on data. Good data documentation is paramount to effective data use. Researchers need metadata or “data about the data” to understand the numbers and transform them into meaningful knowledge. Metadata is essential to evaluate the cross-country comparability of data.
The metadata documentation in Nesstar provides GGS-data users with transparent and easy digestible user-friendly information on data collection and processing procedures. Metadata is provided in compliance with the Data Documentation Initiative (DDI) metadata standard . Through the GGP webpage, users can access a range of relevant metadata on the micro-level datasets, including:
• information about bibliographic citation, study scope and data access;
• information about national data collectors (i.e. producers, distributors, depositors, and funding agencies);
• information on data collection methodology and processing (i.e. sampling procedure, mode of data collection, type of research instrument, characteristics of data collection situation, actions to minimize losses, control operations, weighting, cleaning operations, and response rate);
• overviews of the cross-country availability of variables and descriptions of country-specific deviances from the coding scheme or the target sample;
• information on data entry and data pre-harmonisation by national data collectors, as well as information on central data harmonisation procedures.
Extensive metadata documentation on indicators in the CDB is available as well. Each indicator is linked to a “Meta information sheet”, which includes indicator-level information, such as definition, unit of measurement, time coverage, national data providers, sources used, and the geographical reference areas. In addition, specific metadata on each single data entry is offered. This includes the data source and its web-link, the retrieval date, information on any calculation or deviation from variable definition as well as on any potential break in series (e.g. due to revisions of data collections methods and/ or changes in country boundaries). This ensures a high level of transparency with regard to the origin and quality of the data.
With the implementation of new – simplified – procedures for data access in 2011, data access has been sped up. In the future, the GGP will continue streamlining data access procedures with the goal of releasing downloadable micro-level data to applicants as expediently as possible. The GGP Data Archive will pursue its efforts towards high quality and well-documented data, in compliance with the norms and standards promoted by the archives of the Council of European Social Science Data Archives (CESSDA). This is essential to increase the number of users of GGS micro-data files. In addition, the Nesstar online archive system and the Contextual Database will be maintained and extended, as soon as new data become available.
Stimulating data usage and dissemination of results
Dissemination of social science data and knowledge is rapidly becoming an integral activity of all major Research Infrastructures. The GGP is no exception to this trend, and stimulates the broadest possible usage of GGP data as well as the widest possible dissemination of the results of analyses based on these data to scientific community, the policy community, and the general public. This Section presents an overview of current and planned activities related to the stimulation of data usage and the dissemination of scientific and policy-relevant research results.
Several types of activities have been undertaken to stimulate the use of GGP data. These include (a) raising awareness about the GGP through production of brochures and leaflets, (b) a website with all available information on the GGP, and the option to perform online analysis and to access raw data, (c) the presentation of the GGP at international scientific workshops and conferences, (d) and training activities for researchers. A series of workshops has been organized on issues like data collection, harmonization and archiving, on new measurement instruments, and on new methodological approaches in survey research. In addition, the First GGP User Conference was held in Spring 2011 in Budapest, attracting about 120 researchers. This conference offered the opportunity to researchers to present and discuss their GGP-based research, to learn from each other’s experiences with the GGP data, and to initiate new collaborations.
The GGP Research Infrastructure will continue and deepen its activities to stimulate data usage. The functionality of the Website will be enhanced, with improved information on indicators of data quality and links to key international projects that use GGP data. Several different types of Networking Activities are envisaged as well. Following up on the success of the First User Conference in Budapest, a bi-annual series of User Conferences will be organised to exchange scientific and methodological knowledge and new insights across the GGP community, with the next one taking place in Milano in October 2013. In addition, training workshops and summer schools will be organised to acquaint current and new users with the opportunities that the research infrastructure offers, to facilitate the application of sophisticated methods (multi-level analysis, SEM) to the available data, and to train national teams in the application of commonly agreed upon methodological standards. Depending on funding, we will also develop a GGP Visiting Scholar Programme. Such a Visiting Scholar Programme enables international researchers to spend between two and six months at one or more of the participating institutions in the Programme. The Visiting Scholars could profit from the vast knowledge on the GGP available within these institutions, collaborate in the programme development of the GGP, as well as co-author academic publications with local GGP researchers.
In order to exploit the potential of the GGP, activities will also be undertaken to enlarge the type of datasets that the GGP makes available to its users. These activities will include (a) the preparation of longitudinal data files, event-history data files and other types of integrated data files (e.g. child-oriented data files) that facilitate the full usage of the available data, (b) the harmonisation of non-GGP survey data files (e.g. Understanding Society in the UK) to the GGP Harmonised Data File (HDF) standard, (c) the development of an easy-to-use data file for teaching purposes, both at the Bachelor and the Master level, (d) the preparation of a pooled GGP dataset for online browsing and analysis to be assessed by Nesstar, and (e) the enrichment of the Contextual Database with contextual data based on policy information.
The data provided by the GGP Research Infrastructure are widely used. As the Graph shows, by the end of 2012, more than 680 papers and reports based on GGP data have been published or presented at conferences. This includes more than 200 articles in scientific journals and over 20 PhD theses.
A major aim of the GGP is to disseminate findings based on its data to a wider audience of policy makers as well as to the general public. Since 2012, GGP at a Glance, our bimonthly newsletter is distributed among almost 500 subscribers. Each Newsletter includes an item called Did you know? that alerts the reader to recent scientific and policy-relevant GGP findings. In addition, abstracts are published from articles based on GGP data that have appeared in peer-reviewed journals. Finally, GGP-related news is included in the Newsletter. In addition to the Newsletter, the recently launched GGP Research Notes series contributes to the broader discourse on contemporary population challenges and demonstrates what GGP data can contribute to policy and research agendas. Three to four issues per year will be published, in which internationally renowned scholars discuss GGP research with clear policy implications.
For the future, we aim to organise policy events – both small scale ones, like targeted seminars, and large scale ones, like policy conferences –to disseminate policy-relevant knowledge that is generated by the GGP. The GGP will also work with the United Nations Economic Commission for Europe (UNECE). The UNECE views the GGP as a means to inform policy making about ageing and gender issues. In 2008, a special conference on ‘How Generations and Gender Shape Demographic Change’ was organized by the UNECE to discuss first results based on data from the GGP. Official representatives of 32 UN Member States participated. Dissemination activities will also benefit from the involvement of Population Europe (PE), the collaborative network of leading demographic research institutes and over 200 experts. With support of the European Commission, Population Europe has developed tools to disseminate research outcomes to policy-makers, civil society actors, the media, and lay audiences. Involvement of Population Europe will also enable the use of PE Digest, Population Europe’s series of research briefs that summarize policy-relevant results from scientific research published in international journals. Finally, our aim is to expand our website with GGP material of interest to policy makers.
List of Websites:
The projects website can be found at http://www.ggp-i.org
The Generations and Gender Programme (GGP) is a Social Science Research Infrastructure that provides micro- and macro-level data which significantly improve the knowledge base for social science and policymaking in Europe and developed countries elsewhere. In the period 2009-2012, a Design Study, funded by the EU’s Seventh Framework Programme has been carried out.
The key features that allow the GGP to make a significant contribution to the scientific and policy challenges facing European and other developed societies are:
• Cross-national comparability;
• A longitudinal design, with a three-year interval between survey waves;
• A large sample size, with on average 9,000 respondents per survey wave per country;
• A broad age range, with respondents aged 18 to 80;
• The combination of micro and macro data;
• A theory-driven and multidisciplinary design.
Nineteen countries have conducted at least one wave of the Generations and Gender Survey, while 12 countries have carried out multiple waves. By March 2013, wave 1 micro-datasets for sixteen countries and wave 2 datasets for six countries have been made available to researchers. These data are used in over 500 research projects and have led to over 600 articles, reports and scientific presentations. The GGP Contextual Database provides information for 60 countries on more than 100 indicators.
The Design Study has been used to evaluate the content and quality of the data collection and dissemination procedures and practices. On the basis of that evaluation, plans for the future have been outlined. The most important of these are:
• The development of a questionnaire for the 2015 wave of the Generations and Gender Survey (GGS), showing a well-balanced mix of continuity and innovation.
• The adoption of a mixed-mode survey design to reduce data collection costs, whilst maintaining scientific integrity.
• Regular updating of the Contextual Database, with a focus on increasing the number of available policy indicators.
• Streamlining of data access procedures, to expediently release downloadable micro-level data.
• Sustained efforts to disseminate GGP results to scientific and policy audiences, through our monthly newsletter, a GGP Research Notes series, and by small- and large-scale policy events.
The Design Study has also been used to work towards a sustainable future for the GGP. This includes:
• Increased institutionalization. The GGP undertakes all possible measures to assure inclusion in the ESFRI Roadmap. In addition, the GGP will work towards the establishment of an ERIC.
• Costs reductions. The GGP is committed to run the research infrastructure as cost-effectively as possible. A series of measures to reduce costs have been and will be implemented.
• Fund-raising. The GGP will continue to support fundraising activities in participating countries, but also seek central funding of data collection activities
Overall, the GGP provides a major research infrastructure to inform scientific research and policy debates on issues like intergenerational solidarity, ageing, the nature of family structures, newly emerging social inequalities, work life balance and low fertility; moreover such debates have to be examined in the context of the fiscal crisis and existing policy and structural constraints.
Project Context and Objectives:
In an increasingly globalized world, Europe faces major social, economic, cultural and technological challenges. In Europe 2020, the European Union develops a strategy “to help us come out stronger from the crisis and turn the EU into a smart, sustainable and inclusive economy delivering high levels of employment, productivity and social cohesion”. To realize the smart, sustainable and inclusive growth that the EU envisages, the human capital of Europe’s populations is the continent’s largest asset. To capitalize on that asset asks for a constant reappraisal of how Europeans arrange their lives, both as individuals and as countries. The economic crisis affects not only day-to-day decisions, but also fundamental choices at all stages of people’s lives: marriage and childbearing, the combination of employment and caring responsibilities for the young and the old, retirement, housing, and ageing well. How Europeans cope with these fundamental choices has important consequences for their personal well-being, but also for the adaptability and competitiveness of the societies in which they live. In addition, current demographic developments in Europe have far-reaching implications across all spheres of society, including economic development, social cohesion, sustainability and equity. These developments pose key challenges to public policy-making. The GGP design study aimed to improve our ability to explain the causes and consequences of demographic developments, in particular the changing relationships between genders and generations.
Population challenges. Major demographic developments like increased longevity, low fertility, and high levels of marital instability pose – in particular in combination with other developments like increased female labour force participation – major challenges to European societies. These challenges have become even more urgent given the economic and financial crisis that has struck Europe – and other Western countries – in recent years. Therefore, understanding these challenges and providing answers to them is important both from the perspective of individuals and families, and from the perspective of societies.
Gender relationships. Individuals and couples have aspirations with regard to parenthood and occupational careers and try to strike a balance between the two. At a societal level, a balanced level of fertility and a sufficiently high level of labour force participation are also important to assure a sustainable future.
Intergenerational relationships. People need care. Families are the most important providers of such care. Support is provided both from the older to the younger generation and vice versa. For families, a major challenge is how to uphold the provision of care in situations of increased longevity, smaller families and decreased time availability due to labour market obligations. At the societal level, a challenge is to prepare for a future in which the demand for care is almost inevitably expected to rise.
To improve our understanding of these major societal changes, and their causes and consequences, and to develop appropriate policy responses, top-quality, comparative data are a necessity. In order for the social sciences to provide answers to these questions, data are needed that (1) cover the complete adult life course and focus on key decisions and transitions in men’s and women’s lives, (2) are longitudinal, prospective, and forward-looking, (3) are cross-nationally comparative, and (4) combine information on institutional context and individual behaviour. If we want to prepare Europe for the future ahead, we need data that are based on a life course perspective, which views the course people’s lives take as shaped by earlier life circumstances, the families and social networks in which people are embedded, and the institutional, cultural and economic context.
A breakthrough towards better understanding of the underlying forces behind the demographic developments can be achieved by integrated analyses that address both the societal (macro) level processes and the behavioural mechanisms on the micro level of individuals and households. Such analyses have to rely on appropriately designed surveys that collect micro-level data and on databases including the relevant macro-level data. Much of the necessary macro-level data is available and needs to be adequately integrated into the analysis programmes.
Pertinent differences in long-term demographic development, in the ways these societies are organised, in their cultural characteristics and in their family policies exist across Europe. Disentangling the causes of the differences in demographic reactions requires comparable data from many countries that represent a considerable variety of demographic, social, welfare, and cultural regimes. Such data could only be collected in consolidated international efforts with a high degree of standardization.
Longitudinal panel studies that consider life course experiences from the individuals’ past and relate them to the behaviours that follow can achieve crucial improvements in our understanding of demographic choices and their consequences. To make causal inferences, the analyst needs data where the hypothesized cause is observed before the outcome in a person’s lifetime. The panel design provides a possibility to use the broad range of information pertaining to the time of data collection in explaining subsequent behaviour that is captured in a following panel wave.
The Generations and Gender Programme was initiated in 2000 in response to the data needs identified above. Key features of the GGP initiative are:
• Cross-national comparability. The comparative focus allows analyses of the ways in which policies, culture and economic circumstances influence dependencies between men and women and between the young and the old.
• A longitudinal design. The GGP survey applies a panel design – collecting information on the same persons at three-year intervals – to allow the examination of causes and consequences of inequalities between genders and generations.
• A large sample size. The GGP survey has an average of 9,000 respondents per country, making it possible to study numerical minorities and uncommon events.
• A broad age range. The GGP survey has a focus on the whole adult population and thus includes respondents between the ages of 18 and 80, making it possible to address research and policy issues across the entire life course.
• The combination of micro and macro data. The Contextual Database should include country- and regional level data on demographic, economic, cultural and policy indicators.
• A theory-driven and multidisciplinary questionnaire. The development of the survey is inspired by a theoretical framework that emphasizes that key life-course decisions are influenced by attitudes, social norms and perceived and actual economic and institutional constraints.
Project Objectives
The main objective of the Design Study was to develop a European research infrastructure to ensure a sustainable knowledge base for population related policy formulation. To achieve this main objective, a number of more specific technical and scientific objectives had to be fulfilled. Below, these objectives are outlined.
Providing easy and well documented access to data. Data accessibility is of essential importance for maximising the impact of a Research Infrastructure. The objective of the GGP Design Study was to establish a system that informs the research and policymaking communities of the existence and potential use of the Infrastructure and provides easy access to both the macro and micro level data for the purpose of research. Providing information involves active presentations at appropriate gatherings of researchers and policymakers and inclusion of the relevant information in periodicals and other publications. A user-friendly main web entry point should be designed for a variety of users of this infrastructure, linking to documents of the underlying rationale, on specific technical information as well as to the data itself. Respecting appropriate measures for protecting personal data, appropriate screening and security procedures had to be developed for data access. In this context, the challenge was to ensure unobtrusive and quick data access, which is known to increase the research output based on a data made available by the Research Infrastructure.
Development of new measurement tools. The objective of the GGP Design Study was to improve the validity and reliability of existing measurement instruments and work out ways of collecting the necessary data more efficiently. In many topics envisaged as part of this Research Infrastructure, elaborate specialised measurement tools existed. The objective was to ensure that information allowing highly relevant and innovative analyses combining a number of issues pertinent to demographic behaviour could be collected in the framework of one survey at an acceptable level of respondent burden. Continuous development of measurement instruments contributes to higher data quality as well as to the completeness of gathered information on a specific topic. The measurement instruments should therefore aim towards constant improvement and to utilize new substantive and technical knowledge of the disciplines they are based in. The Design Study aimed at achieving a step-change in the measurement of economic well-being and social support networks as well as decision-making processes in the life course in the context of a large-scale multi-disciplinary survey.
Integrate new research areas. The GGP aims at including all topics with well-established theoretical relevance for explaining demographic behaviours and with a link to public policies. While this greatly facilitated developing survey modules based on a clear theoretical framework, several emerging areas of scientific inquiry into demographic developments had not been covered. Therefore, the Design Study had the objective of broadening the scope of disciplinary perspectives included in the GGP. This would involve developing measurement tools that would allow analysing the interplay of psychological traits, genetic and other biomarkers with demographic behaviours. Emerging research in these areas suggested a potential to contribute greatly towards our understanding of determinants of demographic behaviours and the Design Study integrated leading centres in these research areas, which would be in a position to develop an internationally applicable measurement standard.
Improvement in the quality of measurement instruments and the quality of survey data. In today’s context of decreasing rates of survey participation, securing high quality survey data is becoming more challenging. The objective of the GGP Design Study was to improve the existing GGP survey design to minimise distortions arising at different stages of the data collection process. This should be achieved through research in sampling methods, in methods of improving survey response and in developing data adjustment procedures, applied to the internationally comparable survey with longitudinal design. Continuous development of measurement instruments was expected to contribute to higher data quality as well as to the completeness of gathered information on a specific topic. The measurement instruments should aim towards constant improvement and utilize new substantive and technical knowledge of the disciplines they are based in.
Improvement in the cost-effectiveness of survey data collection. Declining rates of survey participation, increasing cost of personal interviewing as well as the emergence of new technologies drive the need to investigate alternative approaches to data collection. The objective of the Design Study was to improve the quality of the collected data while reducing response burden. Research activities in the field of data collection methods were expected to utilize technological advances such as Computer Assisted Interviewing (CAI) and the opportunities offered by the continuously growing access to the internet. It should furthermore investigate the use of information from administrative registers at different stages such as sample preparation, and data quality control as well as a supplemental source of data. Based on the results, recommendations should be made for the use alternative data collection methods. New findings should also contribute to better understanding of the impact of different modes of data collection and the use of the new technologies on data quality and the results were thus expected to contribute to the survey methodology as a scientific discipline.
Improvement of data comparability across countries. Comparability of data collected across different social, political and economic contexts is a fundamental building block of any cross-cultural study. Only coherent data based on common definitions, classifications and methodological standards can provide sound foundation for comparing different contexts and thus allow making inferences about the size and nature of the contextual effects to inform policy. The objective of the Design Study was to elaborate and implement measures to improve comparability of the internationally collected micro and macro-data. To achieve this, the implementation of the existing data collection instruments in different contexts was to be scrutinized and the collected information analysed.
Development of new methodological approaches in data analysis. The complexity of the integrated design of combining different sources of data as well as different disciplines and measurement instruments is inevitably translated into a complexity of the underlying data structures. The longitudinal nature of the survey design introduces additional considerations such as dealing with missing and incomplete information due to panel attrition. Complex, multilevel data structures require alternative approaches to data analysis. The objective of the Design Study was to investigate the possible approaches towards conceptualizing and testing innovative approaches in analysing incomplete complex data in a longitudinal perspective.
Advance understanding of the survey process. The GGP is the only large-scale multidisciplinary and internationally comparable panel survey, that encompasses the whole key family-building, working and early retirement age range (18-79). As such it presents an immense asset not only for its contributions to the understanding of contemporary demographic challenges, in particular for Europe, but also to possible advancements in survey methods. As a Research Infrastructure, the Design Study aimed to build on methodological excellence of its implementation as well as to contribute new knowledge to the field of survey methodology. The objective of the Design Study is to gain better understanding of the survey process, especially through an inter-cultural perspective, in order to reduce respondent burden and improve the quality of survey data.
Project Results:
In this section, the main scientific results and outcomes will be presented. It is organized around four topics, that reflect both the objectives of the GGP Design Study outlined in the previous section and the organization of the Design Study in a number of Work Packages. These four topics are:
1. Data Availability and Access
2. Quality of Research Instruments
3. Development of New Questionnaire
4. Development of New Data Collection Methods
Data Availability and Access
A key objective of the GGP Design Study was to provide easy and well-documented access to data. Within the Design Study, activities related to data access and documentation were concentrated in WP2 (Data Harmonization), WP3 (Data Repository and Dissemination) and WP4 (Maintenance and Development of Contextual Database). These three WPs were coordinated by the Working Group on Data Preparation and Dissemination.
The GGP Research Infrastructure provides access to micro-level data (collected through the Generations and Gender Survey) and macro-level data. Two types of access to micro-level data are available. The most extensive use – preferred by researchers – is through the analysis of f ully harmonized individual-level datasets that can be downloaded from a secure server by registered and approved users. A more descriptive way of assessing the data – preferred by journalist, policy makers, and researchers orienting themselves about the possibilities of the data – is by analysing the data on-line using the Nesstar analysis tool. Simple registration suffices to use this tool. Access to the macro-level data is through the on-line Contextual Database. In this section, we will describe results and developments concerning:
• Generations and Gender Survey
• Contextual Database
• Data Documentation and Access
Generations and Gender Survey
Micro-level data on intergenerational and gender relationships are collected through the Generations and Gender Survey (GGS). In this Section, we briefly describe the key features of the GGS, the activities undertaken to evaluate the quality of the data collected through the GGS, and the features of the next generation of GGP surveys, titled GGS 2015 and Beyond.
The GGS is a panel survey that focuses on the whole adult, non-institutionalized population, thus allowing it to address research and policy issues across the entire life course. Its longitudinal design, with information on the same persons being collected with a three-year interval, allows the examination of causes and consequences of the transitions individuals experience and the decisions they make. The average number of respondents per wave and per country is 9,000, allowing the GGP to study numerical minorities and uncommon events. Preferably, respondents are interviewed face-to-face, in order to ensure the highest possible level of cooperation. The GGP questionnaire is multidisciplinary and covers a broad array of topics of relevance to demographers, economists, sociologists, social psychologists and epidemiologists. At the same time, the development of the survey has been inspired by a unifying theoretical framework in which key life-course decisions are thought to be influenced by attitudes, social norms, and perceived and actual economic, social and institutional constraints. Key topics covered by the GGS include:
• Fertility
• Partnership
• Transition to adulthood
• Work-family balance
• Gender relations
• Life course and decision-making
• Intergenerational exchanges
• Informal and formal care
• Wellbeing and health
• Grandparenthood
• Economic activity
• Retirement
At the start of the Design Study, GGS data availability was still very limited. For a few countries, non-harmonized wave 1 datasets were available. Within the Design Study, harmonization procedures to ensure the quality and comparability of country-level data were established and refined, first for wave 1 data sets, and later on for wave 2 data sets. To ensure the cross-national comparability of the GGS, transparent and centrally coordinated harmonization activities are essential.
To achieve the highest level of cross-national comparability, the GGP combines both “input harmonization” and “output harmonization”. “Input harmonization” requires country teams that conduct the GGS to implement a centrally developed survey design and model questionnaire. As the details of the survey collection process are largely determined at the country level, and reflect national research interests and financial constraints, the national implementation of the GGS might deviate from the centrally suggested guidelines. These national-level deviations from the central guidelines and the model questionnaire, in combination with the need for rigorous quality control, require data harmonization procedures to be carried out in the post-processing phase. Such “output harmonization” is applied when the national implementation of the survey differs with regard to what is measured and how it is measured.
The initial post-processing activities of data collected by national survey organizations implementing the GGS are conducted by the national GGP teams. Subsequently, in order to assure maximal consistency and the highest possible level of standardization, the final part of the harmonization procedure is carried out centrally by the data harmonization unit of the Netherlands Interdisciplinary Demographic Institute (NIDI). This final part of the harmonization procedure is highly automated using advanced programming techniques in Stata®, a powerful statistical software package with great data manipulation capabilities. These harmonization activities not only check whether countries have consistently implemented the guidelines for pre-harmonization (i.e. variable naming and coding, country specificities, missing data), but also aim to increase the international comparability as well as the user friendliness of GGS data for the scientific community. These procedures include an extensive checking procedure of the routing of the questionnaire and the consistent coding of different types of missing data. Moreover, it encompasses the sorting of grids and event histories, and the construction of variables that contain basic socio-demographic characteristics of the respondent. The harmonization activities conducted at NIDI after the national GGP teams have submitted the pre-harmonized data files are described in detail in Deliverable 30 (Documentation of Harmonization Procedures).
Harmonization is a dynamic process. Since the release of the first harmonized data files in 2009, the scientific community of data users has been keen to communicate its feedback on these files via electronic mail or in personal communication during events such as the GGP User Conferences and other meetings. Moreover, Work Package 2 systematically analyzed the data quality of the GGS data and revealed ways in which the data harmonization procedures could be improved, for instance, in terms of routing checking procedures. Based on the feedback of users and these emerging new requirements, the user-friendliness of GGS data has been increased by evaluating and improving the harmonization procedures. As a result, two major revisions of Wave 1 data have been released during the term of the project. The first revision in October 2010 —Version 3.0—, included a major evaluation and improvement of the routing check procedure that had been moved from an XML environment to a Stata® environment. The second revision in June 2012 —Version 4.0— included a large number of additional constructed variables entailing socio-demographic characteristics of the respondent.
By the end of the Design Study, first-wave data have been made available in a harmonized format for sixteen countries: Austria, Australia, Belgium, Bulgaria, Estonia, France, Germany, Hungary, Italy, Lithuania, the Netherlands, Poland, Romania, Georgia, Norway and Russia. In addition, a dataset of the Turkish sub-sample of the German GGS is available. In 2012, the first second-wave harmonized datasets were released for six countries as well (Bulgaria, France, Germany, Georgia, Hungary, and the Netherlands). Thus, micro-level data on more than 157,000 individual respondents has already being made available.
Contextual Database
From the outset, the GGP has acknowledged that intergenerational and gender relationships are influenced by the demographic, economic, cultural and policy contexts in which individuals are embedded. Thus, a Research Infrastructure that provides scientists, policy analysts and journalists with data to study these issues, should provide both micro- and macro-level data. In line with this fundamental insight, the GGP has developed a Generations and Gender Contextual Database (CDB), next to the national Generations and Gender Surveys (GGS). The CDB serves two main purposes. First, it aims to support research on macro-level influences on micro-level processes, taking both individual-level and contextual-level characteristics into account. Second, it allows researchers to study macro-trends across time and space.
The data for the CDB are derived from two main sources. Where possible, existing online-databases by international organisations and research consortia are used. In addition, in each country participating in the GGP, a group of national experts has been set up to collect very specific indicators at the national and sub-national regional level, which are poorly represented in international databases (e.g. in terms of time coverage). The CDB coordination team carries out the harmonisation process of these data. In this process, each indicator derived from different national and international sources is cross-checked in order to select the best combination of data to be included. The choice of the data sources is determined by the following criteria: (a) compliance with CDB guidelines and international standards, (b) comparability across countries, (c) completeness, spatial and temporal availability of the respective indicator, and (d) availability of well documented metadata information and variable definitions. The process aims at maximising the geographical and temporal coverage of harmonised indicators for which data are made available.
In total, the CDB aims to include more than 200 cross-country comparative indicators, spanning a broad array of 16 different domains. The database includes quantitative indicators which in many cases cover not only national-level data, but also data on the sub-national, regional level. In addition, there are descriptions of main policy changes and reforms over time. In order to comply with the retrospective dimension of the GGS, most indicators cover time-series from 1970 onward or even longer periods of time. At the current stage, contextual data for around 100 indicators have been harmonised and made available in the GGP Contextual Database, covering up to 60 countries in Europe, North America, Asia and Oceania. The major areas covered by these indicators are: Demography, Economy and Social Aspects, Labour, Employment and Pensions, Child Care, Education, Health, Culture, and Tax and Benefits. These indicators can be accessed from the GGP website (http://www.ggp-i.org/) using a user-friendly interface that allows both on-line analysis of trends and easy downloading of indicators in multiple formats.
Overall, the GGP Contextual Database comprises a unique combination of features that might also serve as a blueprint for contextual databases of other surveys. This combination of features includes (1) harmonised cross-country comparative indicators for up to 60 countries combined from different sources, often covering not only the national but also the sub-national regional level; (2) extensive temporal coverage in order to support longitudinal analyses over time; (3) information on policy developments over time; (4) a dynamic user-friendly web environment, which has innovative functionalities both in terms of metadata documentation and of dynamic geo-coding of national as well as regional data in the extraction process. The geo-coding-function aims to support users to link the contextual data to the individual-level data for multi-level analysis. Different geo-code-systems are supported. This includes the system used in the GGS to identify place of residence of an interviewed person, as well as other systems such as NUTS- or OECD-codes. The coexistence of all these features in the GGP-Contextual Database makes it a unique support tool for researchers interested in the micro-macro linkages of social structures and processes.
At the end of the Design Study, the focus of activities is on increasing the number of harmonised policy indicators in the CDB. In this endeavour, close contacts have been established with other database initiatives working on similar aspects. This includes the Comparative Family Policy Database, whose indicators have recently been included in the CDB, the Population and Policy Database of Population Europe and the Multilinks Database on Intergenerational Policy Indicators. Another plan for the future development of the CDB is to include contextual information on changing social norms and attitudes at the national and regional level over time by aggregating response data from surveys such as the European Social Survey and the European Value Survey. In addition, opportunities for cooperation with coordination teams of other surveys (e.g. SHARE, ESS), which also recently started setting up contextual databases, will be explored. The GGP is very open to share its longstanding experience in the development of its GGP Contextual Database with other Research Infrastructures and to cooperate in the collection and dissemination of contextual data, as there is a substantial overlap in the contextual indicators provided by the GGP, ESS and SHARE. We believe that such cooperation will contribute to making the development of these databases more cost efficient and effective.
Data Documentation and Access
Providing easy and well-documented data access is essential to enable the widest possible range of users to benefit from the rich potential of a Research Infrastructure. Therefore, GGP users need quick access to information on the content of the data that the GGP provides, as well as its geographical and temporal coverage. Information on data quality and cross-country comparability has to be easily accessible as well. Quick and high-quality data access can maximise the impact of the GGP both on the research and on the policymaking communities.
GGP survey data and contextual data are made available free of charge through a single online data repository, which can be accessed through the GGP website (www.ggp-i.org). Several data access-options tailored to different audiences are available. Researchers can apply for access to harmonised GGS-data and download these data as soon as their application has been approved and access has been granted. A broader audience is targeted by the Nesstar online system , which allows a wide range of potential users, such as journalists, policy analysts and others, to analyse the survey data on-line by making tabulations and simple analyses. Moreover, the Contextual Database aims to serve both researchers and broader audiences by providing access to macro-level information on demographic, economic and political conditions.
Data access procedures are set up in compliance with Directive 95/46/EC of the European Parliament and of the European Council of 24 October 1995 on the protection of individuals with regard to the processing of personal data and on the free movement of such data. There are three levels of data access. First, full access to harmonized GGS micro-data files is granted to bona fide researchers who register to the online archive for any academic or policy-relevant research question. Second, just by providing a valid e-mail address, any user can gain access to the Nesstar system. Third, the GGP Contextual Database can be accessed without any type of registration.
Countries participating in the GGP provide de-personalised, clean and pre-harmonised individual-level data along with documentation in English for all survey instruments. They also enter into a legally binding agreement with the GGP which allows dissemination of the national surveys by the GGP Data Archive. Data release agreements have been concluded with all countries submitting data. Agreements regulate relevant data protection issues and outline the requirements for potential users who wish to apply for data use.
GGP aims at providing unobtrusive access to micro-data files while at the same time respecting appropriate screening and security procedures for protecting survey data. Since the amendment of the procedure in 2011/12, GGS data from most countries can now be accessed through a new, lighter application procedure where applicants have to sign just one form on “Statement of affiliation, confidentiality and acceptable use for GGS data sets” outlining the obligations to data users. Harmonised micro-data files are released to successful applicants electronically in SPSS and STATA formats. Users are alerted by e-mail about new versions of GGS data files. At the end of 2012, more than 500 projects have been or are using GGP data. The GGP bibliographic database contains information on more than 680 scientific publications that have been produced on the basis of GGS data.
The GGP data archive also provides users with harmonized background documentation on data. Good data documentation is paramount to effective data use. Researchers need metadata or “data about the data” to understand the numbers and transform them into meaningful knowledge. Metadata is essential to evaluate the cross-country comparability of data.
The metadata documentation in Nesstar provides GGS-data users with transparent and easy digestible user-friendly information on data collection and processing procedures. Metadata is provided in compliance with the Data Documentation Initiative (DDI) metadata standard . Through the GGP webpage, users can access a range of relevant metadata on the micro-level datasets, including:
• information about bibliographic citation, study scope and data access;
• information about national data collectors (i.e. producers, distributors, depositors, and funding agencies);
• information on data collection methodology and processing (i.e. sampling procedure, mode of data collection, type of research instrument, characteristics of data collection situation, actions to minimize losses, control operations, weighting, cleaning operations, and response rate);
• overviews of the cross-country availability of variables and descriptions of country-specific deviances from the coding scheme or the target sample;
• information on data entry and data pre-harmonisation by national data collectors, as well as information on central data harmonisation procedures.
Extensive metadata documentation on indicators in the CDB is available as well. Each indicator is linked to a “Meta information sheet”, which includes indicator-level information, such as definition, unit of measurement, time coverage, national data providers, sources used, and the geographical reference areas. In addition, specific metadata on each single data entry is offered. This includes the data source and its web-link, the retrieval date, information on any calculation or deviation from variable definition as well as on any potential break in series (e.g. due to revisions of data collections methods and/ or changes in country boundaries). This ensures a high level of transparency with regard to the origin and quality of the data.
With the implementation of new – simplified – procedures for data access in 2011, data access has been sped up. In the future, the GGP will continue streamlining data access procedures with the goal of releasing downloadable micro-level data to applicants as expediently as possible. The GGP Data Archive will pursue its efforts towards high quality and well-documented data, in compliance with the norms and standards promoted by the archives of the Council of European Social Science Data Archives (CESSDA). This is essential to increase the number of users of GGS micro-data files. In addition, the Nesstar online archive system and the Contextual Database will be maintained and extended, as soon as new data become available.
Quality of Research Instruments
Two key objectives of the GGP Design Study were to improve the quality of measurement instruments in general and the quality of data comparability across countries in particular. Within the Design Study, activities related to these objectives were concentrated in WP6 (Research on Methods of Measurement and Analysis). In this WP, the emphasis was on the reliability and cross-national equivalence of item scales. This WP was coordinated by the Working Group on Methodological Research. In addition, evaluation of questions on specific topics were evaluated within some of the WPs related to Measurement, in particular WP7 (Measuring Economic Well Being), WP8 (Measuring Social Support Networks), and WP9 (Life Course and Decision-Making).
At first sight, the main data quality issues in comparative research do not seem to differ much from those in one-country study. First, there are issues related to quality of the sample that is realized. Is the realized sample a fair representation of the underlying population it is thought to represent? If not, how large is the bias and can we develop methods to correct for the potential biases in the realized sample? Second, there are issues related to the quality of the data collected from the realized sample. Have all the data been collected or are parts missing for a large proportion of respondents? Are the data that have been collected reliable and valid indicators of the concepts that they are expected to measure? However, in comparative research, additional issues arise. It is expected that the data collected in the countries that are incorporated in the data set are comparable in such a way that if differences between countries in an indicator are observed, these differences can be attributed to country-differences in the substantive construct this indicator is expected to represent, rather than to differences between countries in the realized sample or in the reliability or validity of the indicators themselves. And very often, it is unclear whether this expectation is met, as there can be large differences across countries in sampling frames, sampling methods, interviewer modes, interviewer instructions and behaviour, translations of the questionnaire items, etc. This issue of comparative survey quality is of major importance to the GGP, as it is a key research infrastructure for comparative research. But to what extent does the GGS live up to its promises? The data quality of the GGS has been examined with a focus on a variety of aspects of comparative data quality.
With regard to sampling, the use of a probability sample was prescribed by the sampling guidelines, and it turned out that all countries complied to that prescription. The exact methods in which this was done, though, were country-specific. In most countries, a simple random sample was drawn from national registers or Census information, but in countries where this was not feasible, other randomized methods were employed.
Guidelines for GGP fieldwork implementation were not very elaborate, and as a result, most countries adhered to country-specific best practices. Two issues warrant specific attention from GGP users, though. First, the population covered by the survey (in particular the in- or exclusion of the institutional population , and the age range of the sample) differs somewhat between countries. It is suggested that users drop the small numbers of respondents that were outside the defined GGP universe of 18 to 79 years of age or were living in institutions. Second, the timing and duration of the wave 1 fieldwork differed substantially across countries. As a result, country differences could partially reflect period differences that operate across countries. Users should take this into account in interpreting differences between countries.
Response rates in GGP countries varied considerable, from a rather low 42% in Belgium to a very high 97% in Romania. On average, the response rate was about 67%. To put this response rate into perspective, it can be compared to averaged country response rates in other cross-national comparative surveys. The Survey of Health, Aging and Retirement in Europe (SHARE) reports an overall household level response rate for ten countries in wave 1 of 55% . In wave 3 of the European Social Survey (ESS), conducted in 2006, the averaged national response rate for 25 participating countries was 64%. Finally, the 2009 wave of the European Union Survey of Income and Living Conditions (EU SILC) reports a response rate for new households of 73% for 29 participating countries . Although it is hard to draw definite conclusions, given differences in sampling frames, survey mode, and survey agencies (e.g. official statistical offices versus private fieldwork agencies), these figures suggest that the response rate of the GGS compares quite positively with those of other major comparative surveys in Europe.
Even more important than the level of nonresponse, is the bias in the nonresponse. To evaluate that bias, survey data were compared to population data on a range of characteristics. In addition, this comparison was made both for the unweighted sample and for the weighted sample. For unweighted data, considerable bias was detected for such characteristics as age, gender, region, marital status, household size, and level of education. When the data were weighted using either the weights developed by the country teams or the centrally-developed weights, biases with regard to age, gender, region, and household size were substantially reduced. However, biases with regard to marital status and level of education remained. Therefore, users are advised to apply the weights provided in the GGS datasets whenever feasible. At the same time, caution is needed when the data are used to describe marital status and educational distributions of the population at large.
What implications should these findings have for future waves of the GGP? A first implication is to strengthen efforts to increase cross-national comparability. These efforts could include (a) alignment of the timing of waves in the participating countries, (b) more elaborate fieldwork guidelines, and (c) more elaborate and uniform reporting of contact and nonresponse information. Another important implication is to enhance efforts to document the representativeness of the realized samples and to improve uniformity in the weighting procedures. Such efforts could profit from better documentation of national weighting procedures and/or implementation of central weighting procedures.
An important issue in comparative research is the cross-national equivalence of multiple-item sets. Multiple-item sets are mainly used to measure individual scores on theoretical constructs that are not easy to measure, like human values or ethnic stereotyping. Asking multiple questions on the same theoretical construct – a latent variable in the language of Structural Equation Modelling – allows one to test whether individuals answer the same set of questions in the same way. Only if they do so to a considerable extent is it warranted to compare scores across individuals, or – in the case of cross-national equivalence – across countries. In the Design Study the cross-national equivalence of all multiple-items scales included in the GGS has been examined, and it is concluded that only one of them – the De Jong Gierveld loneliness scale – qualifies for the highest level of equivalence – called scalar equivalence – across almost all countries. Many other scales exhibit partial equivalence, which implies that scale scores are not completely comparable across countries. This may sound as ‘bad news’, but (a) the scales in the GGS actually perform quite well compared to scales in other comparative surveys, and (b) often a remedy is available by analyzing GGS data using Structural Equation Modelling (SEM). New developments in SEM now allow multi-country and multi-level analyses, as well as the application of SEM in hazard analyses, increasing its potential for demographers.
Another aspect of data quality is the extent to which all respondents who have to answer a specific question, actually do so. This is usually referred to as the issue of item nonresponse. Item nonresponse can result from the unwillingness of the respondent to provide an answer, for instance to a question he or she experiences as being too private, or from decisions of the interviewer not to pose a required question. In a comparative survey like the GGS, it could also be that country teams that are responsible for the implementation of the GGS decide not to include specific questions or to pose them to a particular selection of respondents only, in that way deviating from the guidelines provided in the GGS model questionnaire. On average, item non response in wave 1 GGS data sets was about 10 per cent. A detailed analysis of specific sub-sections of the GGS questionnaire suggested that decisions by country-teams to pose certain questions to smaller subsets of respondents than proposed by the GGS model questionnaire was a major factor in explaining the high item nonresponse in some GGS sections. In addition, the role played by question characteristics in explaining the item nonresponse patterns has been examined. Most question characteristics did not have strong effects on item nonresponse. Exceptions were questions on other persons (so called proxy questions) and questions about future expectations, that showed a somewhat higher level of item nonresponse.
Special attention has also been paid to assess the quality of specific sets of indicators. One such important set are variables that can be used to measure individuals’ and households’ level of economic wellbeing. It is important to bear in mind that unlike the EU-SILC, the GGS is not designed to provide extensive information about economic wellbeing. Whereas the EU-SILC is the source of information for assessing living conditions in the European Union, the key focus of the GGS is on gender and intergenerational relationships. Still, it has been assessed to what extent – given limited measures –GGS variables reflecting economic wellbeing resemble those of the EU-SILC and hence their usefulness in terms of measuring economic wellbeing for different demographic constellations and different age groups. A key finding is that the GGS with its focus on demographic trajectories and relations between genders and generations offers an important contribution towards assessing the life-course and economic outcomes. Moreover, the GGS is important in the sense that it has a longitudinal design. With consistent measures of economic wellbeing, we are not only able to assess how trajectories may have an impact on current economic wellbeing, but we are also able to assess how demographic changes between waves are related to changes in economic wellbeing. This design will provide important insights that cannot be done with the EU-SILC.
Based on the household income as reported by the respondent, poverty rates were computed by taking a standard approach similar to that used by EUROSTAT using EU-SILC data. Our estimates show that in those countries where household income is given by exact amounts, the estimated poverty rates are very similar to those of EU-SILC. The exceptions are Germany and France, where the GGS poverty rates are somewhat higher. The analysis shows that this is not due to the imputation procedure implemented. Most likely it is a result of the way household income is reported in income bands. The other exception is Hungary where the number of missing values for household income is large. Here the imputation does impact the poverty rates, and our recommendation here is to stick with the original income measure if the aim is to produce reliable poverty statistics. In addition, analyses showed that measures of material deprivation are highly consistent with EU-SILC. Descriptive statistics show that levels of deprivations are consistent with the subjective measures and also estimated poverty rates in that they move in same direction when the levels varies for different household constellations. The measures are also consistent in terms across country levels. France and Germany are the countries with lowest levels of economic deprivation – not matter how it is measured, whereas it is highest for Georgia. Overall, the conclusion is that the economic wellbeing measures in the GGS are of decent quality, but that country differences need to be taken into account when comparative analysis is done. We also feel that the inclusion of these measures provide high value added compared to EU-SILC – especially because of the longitudinal design and its emphasis on demographic processes.
A key innovation of the GGS is its use of the Theory of Planned Behavior (TPB) to guide the selection of questions on intentions concerning major life-events like leaving home, union formation, parenthood and retirement. Questions were not only posed on people’s intentions regarding these behaviors, but also about their attitudes, norms and perceived behavioral control concerning these behaviors. Analysis of the measures included in wave 1 of the GGS showed that the TPB, which forms the basis of the decision-making item sets, is a good framework for continued study of life course decision-making, not only because it is a sound and well-tested theory, but also because it provides a unifying framework for the factors proposed in other literatures to influence life course decision making and it offers valuable guidelines for definition and measurement of cognitions that contribute to life course decisions. It has already been used with success in studies of fertility intentions and realisation of those intentions, particularly those studies conducted as part of the REPRO project. Analyses of the TPB batteries for the other decisions included in the GGS show that the TPB also provides a good explanation of intentions for these actions. Based on the extensive analyses of TPB items, a number of recommendations were made for the development of the new questionnaire. The key recommendation in that respect is the replacement of the entire battery of perceived behavioural control questions for all decisions, in order to directly address the concept of control as defined in the TPB.
The social network indicators in the GGS have also been extensively evaluated. It turned out that the role-relation approach used (i.e. questions ask whether specific types of network members, e.g. partner, children, parents, friends, co-workers, provide and receive specific types of support) worked well, but had one important drawback; it was not possible to link support exchanges to specific members of the network. For instance, if someone has three children, one knows whether these children provide emotional support or not, but not who of these children provide support and who do not. It is suggested to add questions on specific support-providers and receivers in the new GGS questionnaire.
Development of new questionnaire
Many of the activities described in the previous section, directly fed into the development of a new questionnaire for future waves of the Generations and Gender Survey. This new 2015 GGS Questionnaire is one of the key deliverables of the Design Study (D26). A lot of efforts have been put in the preparation of this questionnaire, including:
1. Evaluating the quality of measurement instruments in the GGS wave 1 questionnaire (see the previous section);
2. Assessing the usage and usefulness of different parts of the GGS wave 1 questionnaire;
3. Identifying ‘blind spots’ in our scientific and policy knowledge and how to operationalize these;
4. Testing new questions and different forms of answer formats;
5. Adapting the flow of the questionnaire.
The task of evaluating the previous questionnaire and developing a new questionnaire was a combined effort by, on the one hand, sets of researchers who evaluated the existing parts of the questionnaire from a substantive point of view, and a set of researchers that focussed on the methodological aspects of the questionnaire, on the other hand. Substantively, the development of the new questionnaire focussed on measures of economic well-being (WP7), measures of social support and intergenerational relationships (WP8), measures on key life course decisions, such as leaving home, union formation, parenthood, and retirement (WP9), and measures of psychological traits and social well-being (WP10). Methodologically, the focus was both one the quality of individual measurement instruments, and on the general issue of developing a questionnaire that could be used in a mixed-mode approach to data collection, and yields equivalent responses in face-to-face, Web and telephone modes.
The key decision about GGS 2015 was that the focus should be on continuity. The value of a panel design largely rests on the ability to examined change across personal and historical time, and this is only possible if the same measurement instruments are repeated in subsequent waves of the survey. Thus, measurement instruments were only changed if there were strong methodological reasons to do so. At the same time, some level of innovation is necessary as well, as new topics emerge as being important from either a scientific or policy point of view. Finally, the content of the questionnaire is limited by the length of the questionnaire. It was decided that the length of the GGS 2015 questionnaire should not exceed that of the previous GGS questionnaire. Therefore, space for innovation could only be created by reducing the length of other parts of the questionnaire.
Compared to the previous GGS questionnaire, an important decision was to drop the use of optional modules. In Wave 1, four optional modules (on ‘nationality and ethnicity’, ‘previous partners’, ‘intentions of breaking up’, and ‘housing’) were included in the GGS. However, these modules generally were fielded by only a small number of countries, thus reducing the potential for cross-national comparison.
The new questionnaire was developed within WG3 on questionnaire development. It met four times. The work started with a discussion of the evaluation of the quality of the measurement instruments in Wave 1. Based on this discussion, decisions were made on which parts of the questionnaire needed to be revamped, and which parts could basically be left unchanged. Next, proposals for new questions were developed by the responsible substantive sets of researchers. Based on their proposals, a preliminary version of the questionnaire was prepared and discussed in WG3. Cuts were made to the draft version based on substantive and methodological considerations, but also based on the necessity to limit the duration of the survey. This draft version was discussed with the GGP Consortium Board and the Network of National Focal Points. Next, the draft version was tested in a pilot study in Slovenia. Based on the results of the pilot, the questionnaire was once again discussed in WG3 and adjusted. The adjusted version was then approved in the Consortium Board. The resulting questionnaire is available as D26.
Substantially, the 2015 wave of the GGS has retained its focus on understanding intergenerational and gender relationships. This includes attention to leaving the parental home, the family formation process, the combination of employment and intergenerational care obligations – both towards parents and children, and retirement. At the same time, GGS 2015 contains some innovative aspects. These include:
• A revamping of the social network module. The new social network module makes it possible to more clearly distinguish the support relationships between genders and generations. This is important to better understand why certain family members offer and receive support, whereas others do not.
• Better measures of how societal constraints and social policies influence the decisions that individuals make concerning family formation and retirement. This is important, as one of the main goals of the GGP is to be able to understand how people react to changes in policy environments.
• Inclusion of a personality module. This is important as personality can have an impact on how individuals behave across a wide range of life domains and can have a long-lasting influence on how individuals react to adversity in the lives.
• Revamping of questions on perceived behavioural control regarding leaving home, union formation and parenthood. These questions are part of our Theory of Planned Behaviour approach towards understanding demographic decision-making. However, the questions in the first waves of the GGS did not work as expected, and have been reformulated in collaboration with Icek Ajzen (the designer of the TPB approach).
Development of mixed-mode approach
A central aspect of the design study has also been to evaluate the survey design of the GGS. In general, the survey design of the GGS is excellent, but relatively expensive. The mode of data collection in the first waves of the GGS – face-to-face interviewing – is the gold standard in the field, but is much more expensive than other modes such as telephone or Web. In addition, alternative modes of data collection – and Web-questionnaires in particular – are becoming more popular and allow a substantial reduction in costs. To assess the potential of alternatives to face-to-face, a pilot study was conducted in which alternative data collection systems were compared in terms of data quality and effectiveness. The pilot study was organized in Slovenia. Based on that study and prior considerations, it has been decided to opt for a balanced combination of different survey modes that combine cost-effectiveness and high quality.
The key outcome of that evaluation is the adoption of a mixed-mode survey design, with first-time respondents being surveyed by use of a face-to-face interview, and repeat respondents preferably being surveyed by a web-based questionnaire. The use of web and telephone modes leads to a reduction of overall survey costs. First-time respondents are surveyed via a face-to-face interview as this increases their retention rate. Special care is taken to allow for comparisons of results across different modes.
While a mixed-mode approach will not necessarily be more efficient or cost effective in the initial survey, both the pilot study experiment and previous research indicate that an online access panel strongly suggests a more favourable situation in subsequent waves of GGS. Here, an initial face-to-face contact enables a higher persuasion and legitimacy of the survey. Once this is established and a person accepts to participate in subsequent waves of the survey they will be significantly more willing to participate via a web portal. Moreover, higher response rates to web survey in subsequent waves will greatly reduce overall variable costs of data collection as lower number of face-to-face or telephone interviews will be required. Such reductions in costs can then free up resources for other aspects of the research and hence lower overall costs. Country-specific differences are, however, expected, meaning that cost-reduction levels could differ between countries.
For the longitudinal component of the data collection, it is recommended that the interval between subsequent waves remains three years. This strikes a balance between sample attrition, which would likely be lower with shorter intervals between waves, and the scientific issue of needing enough time for changes in social and demographic behaviour and experience to occur. This, however, requires considerable due diligence in panel maintenance to ensure that attrition is minimized. Here, we recommend the adoption of five practices: a) continuous and close cooperation between the research institute, the fieldwork agency, the interviewers and the respondents; b) incentives for respondents that are of meaningful monetary value and will vary across countries; c) regular contact with respondents through letters, information brochures, requests for updated contact information, and where feasible, the collection of annual information via a short-form questionnaire; d) where possible, interviewer continuity is recommended to help establish a relationship between the respondents and the interview staff; and e) specialized interviewer training and supervision is essential
The overall recommendation is therefore to conduct the first wave with face-to-face survey followed by the web option and starting with the web option in all subsequent waves at three year intervals. If necessary and cost feasible, face-to-face and telephone surveying can be used as alternative modes.
In addition to the general recommendation to move to a mixed-mode survey design within the GGP, a number of additional recommendations were made, including:
• Full computerization of the questionnaire. With computerization, interviewer’s burden decreases, there is less data cleaning, coding and editing in the post-fielding phase, and data quality increases. In addition, centralized computerization of the questionnaire increases control over the questionnaire design and survey implementation in individual countries and lowers costs.
• The GGS keeps its focus as starting from a random sample of the non-institutionalized population, aged 18 till 79, with an average sample size of at least 9,000 respondents per wave. This large sample is necessary because the sample covers the whole adult life span and allows not just for cross-national analysis, but also for analysis at the national and sub-national level.
• The panel character is maintained, but the panel is to be refreshed regularly, to ensure an average sample size of 9,000 per wave. In the refreshment sample, young age groups that were not represented in earlier waves of the GGS, will be oversampled. Weights will be calculated to ensure both longitudinal and cross-sectional representativeness of the resulting dataset.
• The three-year interval between waves is retained as well. This strikes a balance between sample attrition, which would be lower with shorter intervals between waves, and the scientific need for sufficient demographic and socio-economic changes to occur. This three-year interval requires considerable due diligence in panel maintenance to ensure that attrition is minimized. A series of good practices to minimize attrition are proposed.
• A feature of the GGS that is cause for concern is the fact that surveys are conducted at different times in different countries. This lowers the opportunities for cross-country comparison. To increase the temporal comparability, the questionnaire for future waves of the GGP will be regularly updated, and countries are required to field that questionnaire within a limited time-window. In addition, only those country datasets that include at least 70 per cent of the questions in the GGP Standard Questionnaire will be included in the GGP.
Potential Impact
Overall, the GGP is developed to achieve two major types of impact. The first impact is scientific. The combination of surveys and contextual databases provides solid scientific knowledge about how major societal changes shape social and economic inequalities between genders and generations. The key features of the GGP are functional in helping to achieve this impact, such as the inclusion of adults of all ages, the large sample size, the cross-national comparability and the longitudinal focus.
The second impact concerns policy, both at the national and the European level. The GGP is geared towards optimizing not only the quality of empirical knowledge, but also that of policy recommendations. Moreover, the GGP can fill the knowledge gap that exists with regard to the Central and East European countries. By identifying links between policy regimes and generational and gender inequalities in income, health, support, and well-being, the GGP extends the substantive understanding of the risks of social exclusion, social isolation, and poverty. In addition, by organizing meetings with stakeholders and end-users, the infrastructure contributes to the development of new policies that are required to compensate for the potentially negative effects of demographic changes.
Scientific Impact
In this section, we assess the progress made in terms of the generation of new data and the stimulation of new research. We start by describing the numerical increase in data sets, data users and research output. Next, we briefly comment on the content of the generated knowledge.
Exploitation & Data Access
GGP survey data and contextual data are made available free of charge through a single online data repository, which can be accessed through the GGP website (www.ggp-i.org). Several data access-options tailored to different audiences are available. Researchers can apply for access to harmonised GGS-data and download these data as soon as their application has been approved and access has been granted. A broader audience is targeted by the Nesstar online system , which allows a wide range of potential users, such as journalists, policy analysts and others, to analyse the survey data on-line by making tabulations and simple analyses. Moreover, the Contextual Database aims to serve both researchers and broader audiences by providing access to macro-level information on demographic, economic and political conditions.
The first 7 harmonized GGP wave 2 datasets are now available and a total of 16 harmonised wave 1 datasets, with a total of more than 157,000 respondents have been updated or newly released. New data collection efforts have been started in Sweden, while Austria, France, the Netherlands, Italy, Lithuania, Russia, Australia and Hungary have collected subsequent waves of data. GGP data are available to the international scientific community through a single online data repository, which can be accessed through the GGP website (www.ggp-i.org). Over the years, the number of research projects using GGP data has steadily increased, reaching more than 700 in early 2013 (see fig 1.). In addition, the Nesstar online system allows a wide range of potential users, such as journalists, policy analysts and others, to analyse the survey data on-line by making tabulations and simple analyses. The Contextual Database (CDB) that provides access to macro-level information on demographic, economic and political conditions has been expanded and developed. Currently, the CDB includes more than 100 quantitative indicators. In order to comply with the retrospective dimensions of the GGP, most indicators cover time-series from 1970 onward or even further back.
The data provided by the GGP Research Infrastructure are increasingly utilised by the research community. Figure 2 shows the increase in publications based on GGP data. By the end of 2012, more than 1,000 papers and reports based on GGP data had been published or presented at conferences . This includes more than 617 articles in peer-reviewed scientific journals and over 30 PhD theses. The number of conference presentations using the GGP has been growing at a very steep rate and has reached over 450 recorded appearances and is set to grow further when Milan hosts the Second User Conference of the GGP in October 2013. As these numbers reflect, the GGP’s research output is growing and will continue to do so as more longitudinal and comparative data become available.
Existing research based on GGP
GGP data are used by individual researchers studying a specific topic in a specific country, but also by large consortia of institutions and researchers studying a broad subject in a comparative perspective. Below, some of the major projects that have used or are currently using GGP data are briefly described:
• MULTILINKS (How demographic changes shape intergenerational solidarity, well-being, and social integration: A multilinks framework) is an EU-FP7 funded project in which nine scientific institutions collaborated. The objective of MULTILINKS was to investigate how changing social contexts, from macro-societal to micro-interpersonal, affect social integration, well-being and intergenerational solidarity across different European nations. Using GGP data, the project has provided a wealth of new knowledge on intergenerational co-residence, well-being of older adults, and cross-national differences in generational policies.
• REPRO (Reproductive decision-making in a macro-micro perspective) is another EU-FP7 funded project in which ten scientific institutions collaborated to generate new scientific and policy-oriented knowledge on the factors that drive changes in birth rates and influence reproductive decision-making of contemporary Europeans. GGP data were used to study how childbearing intentions and decisions of Europeans were influenced by attitudes, social norms and perceived and actual constraints.
• FamiliesAndSocieties (Changing families and sustainable societies: Policy contexts and diversity over the life course and across generations) is an EU-FP7 project in which 25 institutions collaborate. The project will explore the growing complexity of family configurations and life course transitions across and within European societies and examine their implications for the well-being of children, women and men, inequalities in life chances, intergenerational relations and care arrangements. The project will also investigate how policies address family diversity and its consequences. The aim is to assess the compatibility of existing policies with the increasingly diverse and complex family life courses in Europe, and to contribute to evidence-based policy-making.
• CONOPP (Contexts of Opportunities. Explaining cross-national variation in the links between childhood disadvantage, young adult demographic behaviour and later-life outcomes) is an ERC Advanced Grant project led by Aart Liefbroer at the Netherlands Interdisciplinary Demographic Institute in which GGP data are used to examine the role played by demographic events and trajectories in producing and reproducing inequalities. It studies cross-national variation in the strength of the relationships between childhood social disadvantage, demographic decision-making during young adulthood and subsequent economic, social and health outcomes. The project relies heavily on the use of data from the GGP Research Infrastructure.
• Families in context (Families in context: Unravelling the ways in which policy, economic, and cultural contexts structure generational interdependencies in families and their life outcomes) is an ERC Advanced Grant project led by Pearl Dykstra at Erasmus University Rotterdam. The project’s main objective is to unravel the ways in which policy, economic, and cultural contexts structure intergenerational dependencies in families. Adopting a multigenerational view of family ties, across life phases, it compares and contrasts different theoretical models for why members of different generations help each other.
• FAMHEALTH (Family life courses, intergenerational exchanges and health and well-being in later life), an ERC Advanced Grant project led by Emily Grundy at Cambridge University, will use GGP data, for example, in analyses aimed at uncovering whether country level factors moderate associations between early fertility and later life health and to investigate mid-life (age 35-55) adults’ provision of support to parents and associations with health.
In addition, several other EU-FP6 and FP7 projects, such as MAGGIE, MICMAC and NEUJOBS, as well as a number of projects by ERC Starting (e.g. Arnstein Aassve, Helga de Valk, Letizia Mencarini, Jan van Bavel, Brienna Perelli-Harris) and ERC Advanced Grants (e.g. Gøsta Esping-Andersen) are using or have used data from the GGP.
Policy Impact and Dissemination
A major aim of the GGP is to disseminate findings based on its data to a wider audience of policy makers as well as to the general public. Since 2012, GGP at a Glance, our bimonthly newsletter is distributed among over 500 subscribers. Each Newsletter includes an item called Did you know? that alerts the reader to recent scientific and policy-relevant GGP findings. In addition, abstracts are published from articles based on GGP data that have appeared in peer-reviewed journals. Finally, GGP-related news is included in the Newsletter. In addition to the Newsletter, the recently launched GGP Research Notes series contributes to the broader discourse on contemporary population challenges and demonstrates what GGP data can contribute to policy and research agendas. Three to four issues per year are published, in which internationally renowned scholars discuss GGP research with clear policy implications.
In order to further develop the GGP Community, GGP User Conferences were held in Budapest in 2011 and in Milan in 2013. This 2 day conference saw the presentation of over 50 pieces of original research using GGP data. A 3rd User conference is planned for 2015 and the GGP will encourage both the use of GGP data and the existing findings of GGP research through international conferences and policy networks. As part of these efforts, in 2014 the GGP coordination team will participate in the Berlin Demographic Forum and European Population Conference, providing a bridge between scientific research and policy makers.
In the future, we aim to organise policy events – both small scale ones, like targeted seminars, and large scale ones, like policy conferences –to disseminate policy-relevant knowledge that is generated by the GGP. The GGP will also work with the United Nations Economic Commission for Europe (UNECE). The UNECE views the GGP as a means to inform policy making about ageing and gender issues. In 2008, a special conference on ‘How Generations and Gender Shape Demographic Change’ was organized by the UNECE to discuss first results based on data from the GGP. Official representatives of 32 UN Member States participated. Dissemination activities will also benefit from the involvement of Population Europe (PE), the collaborative network of leading demographic research institutes and over 200 experts. With support of the European Commission, Population Europe has developed tools to disseminate research outcomes to policy-makers, civil society actors, the media, and lay audiences. Involvement of Population Europe will also enable the use of PE Digest, Population Europe’s series of research briefs that summarize policy-relevant results from scientific research published in international journals. Finally, our aim is to expand our website with GGP material of interest to policy makers.
Providing easy and well-documented data access is essential to enable the widest possible range of users to benefit from the rich potential of a Research Infrastructure, including policy makers and other stakeholders. Therefore, the GGP needs to provide quick access to information on the content of the data that the GGP provides, as well as its geographical and temporal coverage. Information on data quality and cross-country comparability has to be easily accessible as well. Quick and high-quality data access can maximise the impact of the GGP both on the research and on the policymaking communities. With our online analysis suite, the GGP has taken the first steps towards this within this design study.
The GGP provides a major research infrastructure that informs policy debates and decisions for the EU and its member states. Among the critical debates requiring this input are intergenerational solidarity, ageing, the nature of family structures, newly emerging social inequalities, work life balance and low fertility; moreover such debates have to be examined in the context of the fiscal crisis, employment patterns, interpersonal support mechanisms, and existing policy and structural constraints. A properly supported, sustainable GGP is of crucial importance for both social science research and policy formulation for Europe and its institutions.
Potential Impact:
The main aim of the GGP Research Infrastructure has been and will remain to provide social science data that can improve our scientific understanding of key societal challenges and to disseminate that knowledge to the scientific public via books and articles in scientific journals and to policy makers and the general public through newsletters and presentations at policy-related meetings. Below, we will first outline the kind of challenges to which the GGP provides answers, and how the GGP relates to major international research agendas. Next, we will provide an overview of the main activities that have been conducted during the FP7 project. Finally, we will outline plans for the future.
Societal challenges
Although the current economic crisis has led to a sharp increase in unemployment, long-term trends still indicate the emergence of significant labour force shortages. One way of resolving labour market shortages is increasing the labour force participation of women. To do so, the relationship between work and family life must be considered and implies a fresh look at how women and men divide tasks in the employment and family domains. The GGP contains information on educational and work careers, division of labour within the household and on preferences of men and women about this division of labour, that allow us to explore the coherence of career choices and the impact of macro factors such as the financial crisis, but also micro factors such as relationship developments, the arrival of a new child or the emergence of care requirements of dependent relatives. For example, the GGP helps us understand the gender specific challenges of employment and career advancement of highly educated and skilled women such as care arrangements and career breaks.
Another important issue is to increase the labour market participation of older workers. The GGP covers the whole life span and enables specific analysis of older workers within the economy. Specific questions regarding retirement intentions are included within the questionnaire and the longitudinal nature of the dataset allows for comprehensive mapping of retirement plans. Detailed information on the nature of work such as the type of contract, supervisory responsibilities, flexible hours and capacity to work from home are all included in order to enable policy makers and researchers an opportunity to understand the strategies for managing care responsibilities and the transition to retirement.
Population ageing has led to discussions on how intergenerational solidarity between younger and older generations can be maintained. It also has led to discussions about increasing demands on adult children to care for their frail parents. The GGP allows us to examine issue of support exchange not just from the point of view of the older generation, but from that of the younger generation as well. It will also allow us to examine under which conditions support exchange between parents and children has positive consequences for the well-being of support provider and support recipient, and under which conditions these consequences are negative.
The process of globalization has led to an emphasis on the flexibilisation of the labour market. This has weakened the position of young adults, a process that is exacerbated by the current economic crisis. As a result, social inequalities in the Netherlands and across Europe might increase. What will be the short- and the long-term impact of the recent (and in some countries continuing) financial crisis on individuals? In particular, and considering that the youth unemployment rate in the Eurozone stands currently at 24%, what will be its impact on the transition to adulthood, that is, on young people’s ability and decision to form an independent household, form a stable partnership, and enter parenthood? And how will the lives of these young adults compare to those of their parents? And how will they differ for men and women? The GGP can provide answers to these questions.
The process of ageing, and the concurrent partial retreat of the welfare state, makes the issue of healthy ageing a pertinent one. The GGP starts from the assertion that understanding ageing issues asks for a focus on the whole life course (what are the precursors of health in old age?), needs attention to the social networks surrounding people (who cares for whom?), and requires a broad conceptualization of well-being (social, economic and medical aspects need to be included). For example, loneliness has long been a topic that has featured heavily within the research outputs of both the NKPS and the international GGP . For society as whole the value of this research is also in the identification of individuals lacking in such relationships and how they might be supported in times of need and support.
The GGP also enables high quality policy research and analysis that are essential for efficient and effective government. Firstly, the multilevel nature of the GGP, enables for the reconceptualization of the welfare state using comparative data at the macro and micro level. Using the Contextual Database which includes policy information for all GGP countries, it is possible to map constellations of welfare arrangements across 19 countries. This will allow for a historical understanding of the Netherlands positioning within past and future welfare state constellations and the ability to contextualize the micro level survey data within these developments. In addition to this the GGP is capable of facilitating evaluations across a broad range of social and economic policies. At a time of austerity and in a climate where efficiencies are a priority, the crucial question for government and civil society is which policies work and which don’t. The detailed policy information available from the registers provided by CBS will enable the GGP to be used in complex policy analysis. For example, it will be possible to follow care arrangements for infirm spouses as the policy mix for elderly care changes in nature. Employment, housing, health & care, social insurance and education are all policy areas that could benefit from analysis of GGP data as it provides an economic and scientifically sound alternative to the gold standard of experimental policy studies.
Links to key policy and research agenda’s
Europe 2020, the vision document published by the Commission, is the European Union’s growth strategy for the coming decade. In a changing world, the EU aims to become a smart, sustainable and inclusive economy based on high levels of employment, productivity and social cohesion. The vision document notes that a successful EU 2020 strategy must be built on a good analysis of the constraints facing policy makers and acknowledges the challenge presented by European demography. The changing demographic landscape, with a lower proportion of young people in the overall population, is expected to result in substantially reduced potential growth by 2020. The crisis has exacerbated the long-term social challenges that Europe faces today, such as solidarity between generations in the context of an ageing society, social exclusion and child poverty, and the integration of an increasing immigrant population. Europe 2020 emphasizes that new policies must demonstrably contribute to social cohesion, tackling unemployment and fostering social inclusion of young and old, and men and women. The EU 2020 strategy is also is an agenda for all Member States, large and small, old and new, highly developed and still developing. Based on Europe 2020, the Commission has developed its research framework programme for the period 2013-2020, called Horizon 2020. Some of the key societal challenges that Horizon 2020 distinguishes are ‘Health, demographic change and wellbeing’, ‘Europe in a changing world: inclusive, innovative, and reflective societies’, and ‘Secure societies’. With its wide coverage of European countries and focus on the relations between generations and genders, the GGP is ideally suited to address these social challenges.
Population ageing has been a major focus within the United Nations Economic Commission for Europe (UNECE). In the Madrid International Plan of Action on Ageing (2002), and subsequent the Ministerial Conferences on Ageing (2008 and 2012), several key commitments and action strategies have been formulated. Advancing gender equality and the empowerment of women became a goal of the United Nations shortly after the Organisation was established. The first meeting of the UN Commission on the Status of Women was held in 1956. More recently, the vital importance of a life-course approach and of cross-nationally comparable survey data was stressed in the UNECE’s Chair summary of the 2013 Regional Conference on population and development . In particular, a life-course approach was deemed necessary to better understand how people’s lives’ unfold as they move in and out of partnership, as they experience various family transitions, and as they are confronted by unequal gender norms. Thus, it should come as no surprise that the UN has been the main driving force in establishing the GGP. They view the International GGP as a means to inform policy making about ageing and gender issues (see the Letter of support by UNECE).
Main activities during the FP7 project and future outlook
Providing easy and well-documented data access is essential to enable the widest possible range of users to benefit from the rich potential of a Research Infrastructure. Therefore, GGP users need quick access to information on the content of the data that the GGP provides, as well as its geographical and temporal coverage. Information on data quality and cross-country comparability has to be easily accessible as well. Quick and high-quality data access can maximise the impact of the GGP both on the research and on the policymaking communities. This Section describes how the GGP assures this kind of easy data access, and outlines plans for the future.
Data access
At the end of 2012, more than 500 projects have been or are using GGP data. The GGP bibliographic database contains information on more than 680 scientific publications that have been produced on the basis of GGS data.
GGP survey data and contextual data are made available free of charge through a single online data repository, which can be accessed through the GGP website (www.ggp-i.org). Several data access-options tailored to different audiences are available. Researchers can apply for access to harmonised GGS-data and download these data as soon as their application has been approved and access has been granted. A broader audience is targeted by the Nesstar online system, which allows a wide range of potential users, such as journalists, policy analysts and others, to analyse the survey data on-line by making tabulations and simple analyses. Moreover, the Contextual Database aims to serve both researchers and broader audiences by providing access to macro-level information on demographic, economic and political conditions.
Main Nesstar interface: Example
Data access procedures are set up in compliance with Directive 95/46/EC of the European Parliament and of the European Council of 24 October 1995 on the protection of individuals with regard to the processing of personal data and on the free movement of such data. There are three levels of data access. First, full access to harmonized GGS micro-data files is granted to bona fide researchers who register to the online archive for any academic or policy-relevant research question. Second, just by providing a valid e-mail address, any user can gain access to the Nesstar system. Third, the GGP Contextual Database can be accessed without any type of registration.
Countries participating in the GGP provide de-personalised, clean and pre-harmonised individual-level data along with documentation in English for all survey instruments. They also enter into a legally binding agreement with the GGP which allows dissemination of the national surveys by the GGP Data Archive. Data release agreements have been concluded with all countries submitting data. Agreements regulate relevant data protection issues and outline the requirements for potential users who wish to apply for data use.
GGP aims at providing unobtrusive access to micro-data files while at the same time respecting appropriate screening and security procedures for protecting survey data. Since the amendment of the procedure in 2011/12, GGS data from most countries can now be accessed through a new, lighter application procedure where applicants have to sign just one form on “Statement of affiliation, confidentiality and acceptable use for GGS data sets” outlining the obligations to data users. Harmonised micro-data files are released to successful applicants electronically in SPSS and STATA formats. Users are alerted by e-mail about new versions of GGS data files.
Data documentation
The GGP data archive also provides users with harmonized background documentation on data. Good data documentation is paramount to effective data use. Researchers need metadata or “data about the data” to understand the numbers and transform them into meaningful knowledge. Metadata is essential to evaluate the cross-country comparability of data.
The metadata documentation in Nesstar provides GGS-data users with transparent and easy digestible user-friendly information on data collection and processing procedures. Metadata is provided in compliance with the Data Documentation Initiative (DDI) metadata standard . Through the GGP webpage, users can access a range of relevant metadata on the micro-level datasets, including:
• information about bibliographic citation, study scope and data access;
• information about national data collectors (i.e. producers, distributors, depositors, and funding agencies);
• information on data collection methodology and processing (i.e. sampling procedure, mode of data collection, type of research instrument, characteristics of data collection situation, actions to minimize losses, control operations, weighting, cleaning operations, and response rate);
• overviews of the cross-country availability of variables and descriptions of country-specific deviances from the coding scheme or the target sample;
• information on data entry and data pre-harmonisation by national data collectors, as well as information on central data harmonisation procedures.
Extensive metadata documentation on indicators in the CDB is available as well. Each indicator is linked to a “Meta information sheet”, which includes indicator-level information, such as definition, unit of measurement, time coverage, national data providers, sources used, and the geographical reference areas. In addition, specific metadata on each single data entry is offered. This includes the data source and its web-link, the retrieval date, information on any calculation or deviation from variable definition as well as on any potential break in series (e.g. due to revisions of data collections methods and/ or changes in country boundaries). This ensures a high level of transparency with regard to the origin and quality of the data.
With the implementation of new – simplified – procedures for data access in 2011, data access has been sped up. In the future, the GGP will continue streamlining data access procedures with the goal of releasing downloadable micro-level data to applicants as expediently as possible. The GGP Data Archive will pursue its efforts towards high quality and well-documented data, in compliance with the norms and standards promoted by the archives of the Council of European Social Science Data Archives (CESSDA). This is essential to increase the number of users of GGS micro-data files. In addition, the Nesstar online archive system and the Contextual Database will be maintained and extended, as soon as new data become available.
Stimulating data usage and dissemination of results
Dissemination of social science data and knowledge is rapidly becoming an integral activity of all major Research Infrastructures. The GGP is no exception to this trend, and stimulates the broadest possible usage of GGP data as well as the widest possible dissemination of the results of analyses based on these data to scientific community, the policy community, and the general public. This Section presents an overview of current and planned activities related to the stimulation of data usage and the dissemination of scientific and policy-relevant research results.
Several types of activities have been undertaken to stimulate the use of GGP data. These include (a) raising awareness about the GGP through production of brochures and leaflets, (b) a website with all available information on the GGP, and the option to perform online analysis and to access raw data, (c) the presentation of the GGP at international scientific workshops and conferences, (d) and training activities for researchers. A series of workshops has been organized on issues like data collection, harmonization and archiving, on new measurement instruments, and on new methodological approaches in survey research. In addition, the First GGP User Conference was held in Spring 2011 in Budapest, attracting about 120 researchers. This conference offered the opportunity to researchers to present and discuss their GGP-based research, to learn from each other’s experiences with the GGP data, and to initiate new collaborations.
The GGP Research Infrastructure will continue and deepen its activities to stimulate data usage. The functionality of the Website will be enhanced, with improved information on indicators of data quality and links to key international projects that use GGP data. Several different types of Networking Activities are envisaged as well. Following up on the success of the First User Conference in Budapest, a bi-annual series of User Conferences will be organised to exchange scientific and methodological knowledge and new insights across the GGP community, with the next one taking place in Milano in October 2013. In addition, training workshops and summer schools will be organised to acquaint current and new users with the opportunities that the research infrastructure offers, to facilitate the application of sophisticated methods (multi-level analysis, SEM) to the available data, and to train national teams in the application of commonly agreed upon methodological standards. Depending on funding, we will also develop a GGP Visiting Scholar Programme. Such a Visiting Scholar Programme enables international researchers to spend between two and six months at one or more of the participating institutions in the Programme. The Visiting Scholars could profit from the vast knowledge on the GGP available within these institutions, collaborate in the programme development of the GGP, as well as co-author academic publications with local GGP researchers.
In order to exploit the potential of the GGP, activities will also be undertaken to enlarge the type of datasets that the GGP makes available to its users. These activities will include (a) the preparation of longitudinal data files, event-history data files and other types of integrated data files (e.g. child-oriented data files) that facilitate the full usage of the available data, (b) the harmonisation of non-GGP survey data files (e.g. Understanding Society in the UK) to the GGP Harmonised Data File (HDF) standard, (c) the development of an easy-to-use data file for teaching purposes, both at the Bachelor and the Master level, (d) the preparation of a pooled GGP dataset for online browsing and analysis to be assessed by Nesstar, and (e) the enrichment of the Contextual Database with contextual data based on policy information.
The data provided by the GGP Research Infrastructure are widely used. As the Graph shows, by the end of 2012, more than 680 papers and reports based on GGP data have been published or presented at conferences. This includes more than 200 articles in scientific journals and over 20 PhD theses.
A major aim of the GGP is to disseminate findings based on its data to a wider audience of policy makers as well as to the general public. Since 2012, GGP at a Glance, our bimonthly newsletter is distributed among almost 500 subscribers. Each Newsletter includes an item called Did you know? that alerts the reader to recent scientific and policy-relevant GGP findings. In addition, abstracts are published from articles based on GGP data that have appeared in peer-reviewed journals. Finally, GGP-related news is included in the Newsletter. In addition to the Newsletter, the recently launched GGP Research Notes series contributes to the broader discourse on contemporary population challenges and demonstrates what GGP data can contribute to policy and research agendas. Three to four issues per year will be published, in which internationally renowned scholars discuss GGP research with clear policy implications.
For the future, we aim to organise policy events – both small scale ones, like targeted seminars, and large scale ones, like policy conferences –to disseminate policy-relevant knowledge that is generated by the GGP. The GGP will also work with the United Nations Economic Commission for Europe (UNECE). The UNECE views the GGP as a means to inform policy making about ageing and gender issues. In 2008, a special conference on ‘How Generations and Gender Shape Demographic Change’ was organized by the UNECE to discuss first results based on data from the GGP. Official representatives of 32 UN Member States participated. Dissemination activities will also benefit from the involvement of Population Europe (PE), the collaborative network of leading demographic research institutes and over 200 experts. With support of the European Commission, Population Europe has developed tools to disseminate research outcomes to policy-makers, civil society actors, the media, and lay audiences. Involvement of Population Europe will also enable the use of PE Digest, Population Europe’s series of research briefs that summarize policy-relevant results from scientific research published in international journals. Finally, our aim is to expand our website with GGP material of interest to policy makers.
List of Websites:
The projects website can be found at http://www.ggp-i.org
