All time references are in CEST
Innovating social research by using existing survey data to build new databases 2
| Mr Christof Wolf (GESIS Leibniz-Institute for the Social Sciences)
Ms Irina Tomescu-Dubrow (Institute of Philosophy and Sociology, Polish Academy of Sciences )
|Wednesday 19 July, 16:00 - 17:30
Survey research in the social sciences can look back at more than 80 successful years. During this period survey data have become one of the most important sources of data for social research. Increasingly, these data are available for secondary research and offer great opportunities for combining and cumulating different sources to cover longer time periods and/or different countries. The modern history of many (western) countries can now be based on these data reflecting the development of social differentiation and social structure as well as changing values, attitudes and behaviors.
This session invites presentations that exploit the value of combined data sources to study social change or cross-national comparisons. We also invite presentations dealing with the methodological or practical challenges faced by projects striving to combine survey data for secondary usage, as well as challenges of analyzing combined data sources. These challenges might include questions of finding and accessing data, variability in data quality, the availability of relevant documentation, the identification of “identical” measures across different sources, how to weight respondents' answers in combined data sources, etc.
Keywords: Survey Data Recycling, Cumulation, Linking, Harmonization
Ms Myriam Baum (Federal Institute for Vocational Training and Education (BIBB)) - Presenting Author
Ms Marion Thiele (Federal Institute for Vocational Training and Education (BIBB))
Dr Dominik Becker (Federal Institute for Vocational Training and Education (BIBB))
Professor Harald Pfeifer (Federal Institute for Vocational Training and Education (BIBB))
Both the business cycle (BC) and technological change heavily affect the labour market’s skill demands. Further vocational training (FT) helps employees to deal with those changes and ensures employability. However, only limited research on this topic exist. One reason for this research gap is the lack of an encompassing data-base. Most research is restricted to major events, e.g. the financial crisis 2008. Following the human capital approach, in a recession individuals might invest more in individual FT to substitute lacking firm sponsoring, but increased financial constrains might hinder individuals’ investments. Next to the BC, technological change and associated task changes affect individuals’ participation in FT. As those two mechanisms are strongly interrelated, we analyse - as a part of a larger project on the BC and FT – whether technological change moderates the relation between the BC and individual FT.
Therefore, we create a new data-base by using the adult cohort (SC6) of the National Educational Panel Study (NEPS; 2007-2020 + Covid survey) enriched by both administrative data of various BC indicators (e.g. unemployment rates; GDP) and firm-level data on technology indicators (e.g. Mannheim Innovation Panel). BC indicators are merged to NEPS three months prior to individuals’ participation in FT, or in case of no FT 15 months prior to the interview. For the linkage of the administrative data we use regional, year and sectoral information; for the linking of firm-level data we additionally use firm-size information. For our analysis we will use panel regressions that address additional methodical problems like unobserved heterogeneity and reversed causality. We will present how we created this unique data-set, which helps both to understand individuals’ FT investments in the times of economic crisis and technological change, and to facilitate applied researchers’ methodological challenges of data integration.
Professor Zbigniew Sawiński (Institute of Philosophy and Sociology Polish Academy of Sciences) - Presenting Author
Professor Anna Kiersztyn (Department of Sociology University of Warsaw)
Dr Katarzyna Kopycka (Department of Sociology University of Warsaw)
Scholars agree that measurement errors should be controlled when integrating data from different survey projects. In this presentation, we deal with the recall bias which arises when data is collected using retrospective questions. The frequency and magnitude of memory errors generally depend on the time between the interview and the event in question. To assess the impact of memory errors on survey results, we propose the Retrospective Distance Scale (RDS), which reflects the time distance between the event under study and the interview. The RDS method provides a test of whether the memory errors are statistically significant, and if so, then an estimation of the magnitude of these errors is provided. On this basis, the recall bias can be at least partly removed, for example, by the imputation of values not affected by memory errors.
The presented approach requires that the RDS values and the respondent's characteristics, including age, are not collinear. This condition is met in most longitudinal studies, except some cohort studies.
The advantages of the RDS method are illustrated with a comparison of the Polish Panel Survey POLPAN, conducted every five years, and the National Longitudinal Survey of Youth 1997, in which the interviews were conducted annually until 2011 and every two years since then. The comparison confirms that panel surveys using different interview cycles may differ according to the recall bias. When harmonizing panel survey data from different countries, the RDS method can help to control the consequences of recall bias resulting from country-specific survey designs.
The RDS values do not require mapping to data before survey dissemination but can be prepared and shared by data users.
Mr Dennis Oliver Kubitza (Federal Institute for Vocational Education and Training (bibb)) - Presenting Author
The transition to universities imposes multiple decisions upon adolescents: choice of subject, university, moving or commuting and related to that, the finance of their lifestyle between leisure and learning. We know that the willingness to commute and move differs dependent on the socio-economic background and regional contexts. Missing models and heterogeneous effects make it difficult in this setting to estimate relevant policies for reducing inequality in opportunities of individuals: Are current subsidiaries of public transport enough to reduce inequality? Are housing subsidiaries an effective alternative to commuting subsidiaries and for whom?
A recent approach to explain these group differences is the acceptable travel time (ATT) as a behavioural threshold (Milakis et al, 2015). Our goal is to identify and estimate differences in ATTs for University students in Germany, by investigating the binary movement decisions of university students in dependence on their ex-ante commuting times. Our multi-level logistic regression models are based on individual data from the MESARAS survey (Weisser, 2016) (N=2589). For assessing the different dimensions of influences, we use that personal information and link these with the regional contexts: average rent prices, labour market conditions and availability of public transport subscriptions for students. Furthermore, we use the geo-referenced mobility episodes (provided by Mesaras) to collect and include commuting times by different modes of transportation (Google Distances API) to generate our key dependent variable.
Weisser, R. (2016). Mesaras 2013: Mobility, expectations, self-assessment and risk attitude of students.
Milakis, D., Cervero, R., van Wee, B., & Maat, K. (2015). Do people consider an acceptable travel time?evidence from berkeley, Journal of Transport Geography,