All time references are in CEST
Advances in the integration of geospatial data for social research |
|
| Session Organisers | Dr Dennis Abel (GESIS - Leibniz Institute for the Social Sciences) Dr Stefan Jünger (GESIS - Leibniz Institute for the Social Sciences) Mr Jonas Lieth (GESIS - Leibniz Institute for the Social Sciences) Ms Anne-Kathrin Stroppe (GESIS - Leibniz Institute for the Social Sciences) |
| Time | Tuesday 18 July, 09:00 - 10:30 |
| Room | |
Data integration of social indicators from surveys and social media content with geospatial context variables has rapidly progressed in the last years. Applications in the social sciences are manifold and cover issues such as conflict and migration, political participation, environmental attitudes, and inequality. Geospatial approaches allow researchers to introduce new perspectives in explaining societal processes, stress the local in questions of global relevance, and extend the potential of inter- and transdisciplinary research. To further progress the data integration based on spatial identifiers, advances are required in several key aspects: 1. High-quality georeferenced survey datasets are needed to expand the pool of linkable social indicators; 2. Social scientists require easier access to georeferenced context variables; 3. Similarly, linking processes need to be simplified and standardized. These steps are crucial for this line of research but bring their own challenges in terms of data quality and fit-for-purpose of the respective datasets to be linked, as well as pitfalls in the linking process itself. Therefore, this session will provide a platform to address these potentials as well as barriers in detail. It collects contributions that aim to advance geospatial social research and is open to empirical, methodological, data-focused, and conceptual papers. Possible topics include (but are not limited to):
- New methods and applications of linking survey and spatial data
- Utilization of novel sources of spatial data to integrate into survey research and answer social science questions
- Quality assurance of the data integration process and resulting data
- Applications and real-world examples integrating survey data with big spatial data, earth observation data, or other small-scale spatial data
Keywords: geospatial data, data integration, data linkage
Dr Sebastian Bähr (Insitute for Employment Research (IAB)) - Presenting Author
Dr Jonas Beste (Insitute for Employment Research (IAB))
Dr Corinna Frodermann (Insitute for Employment Research (IAB))
Dr Jan Mackeben (Insitute for Employment Research (IAB))
Mr Benjamin Krebs (Vienna University of Economics and Business (WU Wien))
While methodological research has identified various predictors of telephone survey response, including respondent characteristics, social environment, interviewer traits, and survey design, some factors remain understudied. One such overlooked but signifi-cant variable is the weather. Our research introduces a novel perspective by exploring the influence of weather on daily activities and time allocation at home. Specifically, we examine how favorable weather conditions, such as higher temperatures, create opportu-nities for outdoor activities.
Consequently, weather conditions may impact the perceived opportunity costs of survey participation. Contact attempts might be more successful in unfavorable weather than in favorable conditions. Our findings could change how we understand and conduct tele-phone surveys. While previous studies have explored weather effects, they used too coarse weather data, focused only on realized interviews, or failed to control for con-founding influences.
Our study addresses this gap by examining how local weather conditions during contact attempts affect the probability of successful contact and cooperation in telephone inter-views. We test whether samples from different weather conditions differ in sociodemo-graphic composition. Using a unique longitudinal dataset that integrates detailed contact data from a large-scale German panel study with fine-grained spatial and temporal weather data, we rigorously test and rule out alternative explanations for weather ef-fects. In line with our opportunity cost argument, we find higher contact and coopera-tion probabilities in cold weather, while both drop in hot weather. Weather-related non-response varies by age but not gender. Our results contribute to the literature on survey response predictors and could enhance telephone data collection efficiency.
Mr Daniel Hein (German Centre for Higher Education Research and Science Studies) - Presenting Author
Many studies in the field of higher education use distance as a simple measure of accessibility, commuting, or moving. They often define distance as the straight-line distance, while only a few studies measure distance using the actual travel distance. This approach is supposedly more sophisticated, accurate, and realistic. Our aim is to assess whether the straight-line is an adequate proxy for travel distances by car and multimodal public transportation in Germany.
We compare the straight-line and the travel distances between the former school and the current university. We also distinguish between the shortest and the best route. The linear relation between the straight-line and travel distance was analyzed using ordinary least-squares regression. To examine outliers, the difference between the actual travel distance and the predicted travel distance, which is the straight-line distance multiplied by the regression slope, was used. The straight-line distance is a good proxy when the absolute difference between the actual travel distance and the predicted travel distance is less than 5 km, or the relative difference is less than 10 %. The results are based on a representative sample of 2,903 different routes taken by German students.
In 96 % of the cases, the straight-line distance is an adequate proxy for the shortest travel distance by car. However, the straight-line is a good approximation of the best car route 80 % of the time. For the shortest and best public transportation routes, the straight-line is a reliable proxy 66 % and 60 % of the time, respectively. The largest discrepancies occur in areas with physical obstacles such as lakes, rivers, mountains or wilderness and nature conservation areas. These findings suggest that future studies should use travel distances for more realistic results, as they provide significantly greater accuracy than straight-line distances.
Ms Ruth Overmann (University of the Bundeswehr Munich) - Presenting Author
Professor Tim Williams (University of the Bundeswehr Munich)
Reproductive decisions are sensitive to objective and subjective stressors that reduce the stability and predictability of life circumstances, as previous research has shown. The perception of and coping with uncertainty depends not only on personal characteristics, but also on the political, cultural and social contexts in which individuals are embedded. In the midst of increasing global uncertainty, cross-national research on the impact of crises on population dynamics is therefore of high relevance.
Violent conflict represents an unparalleled source of uncertainty, and its detrimental effects on populations have been well studied. Yet, studies focusing specifically on fertility intentions are limited. This study aims to further investigate the vulnerability of childbearing desires in armed conflict settings using harmonized data on women’s short-term fertility intentions in over 60 countries from 1988 to 2023.
We combine nationally representative micro-level data from the Demographic and Health Surveys with high-quality georeferenced event data from the Uppsala Conflict Data Program including information on the type of government from the Variety of Democracy dataset to perform geospatial analysis. The analysis is limited to married women aged 15-49 years. Multilevel modelling is applied for an in-depth consideration of country characteristics, allowing for an adequate analysis of the temporal scope and decisiveness of fertility intentions. Conflict events are differentiated according to the duration, intensity, type and geographical proximity of the conflict. We account for a range of control variables.
We expand previous research on the relationship between the exposure to armed conflict and fertility preferences by concentrating on (a) short-term fertility intentions rather than long-term fertility ideals and (b) general effect mechanisms that transcend specific conflict contexts. This focus helps to shed light on the impact of structural constraints on the formation and adaption of childbearing plans.
Professor Tobias Rüttenauer (University College London) - Presenting Author
Dr Kasimir Dederichs (Nuffield College)
Dr David Kretschmer (Nuffield College)
Residential assimilation, or the mutual exposure of majority and minority groups, is essential for immigrant incorporation. While high levels of immigrant-native segregation can support the development of cultural infrastructure within specific neighbourhoods, they often impede direct intergroup contact, leading to perceived threats and diminished mutual understanding. Therefore, understanding the patterns and drivers of residential segregation is crucial. One approach to analysing the drivers of residential segregation is by comparing spatial segregation patterns across different cities. However, this has been challenging due to inconsistent group definitions of minority groups, different spatial scales and data availability across countries.
This study examines residential segregation of immigrants in urban areas using grid-based neighbourhood definitions and harmonized census information. We overcome previous limitations by using grid-based neighbourhood definitions and harmonized census information across 27 European countries and harmonised data from England. Our aim is twofold. First, we present descriptive results on the levels and spatial patterns of ethnic residential segregation across Europe. Second, we investigate the socio-economic correlates of ethnic segregation by merging our segregation indices with additional city-level socio-economic data and employing multi-level models.
We have just received the initial descriptive results and are in the process of merging additional city-level characteristics. I'm sure that we will have the full results by the time of the conference.
Mrs Kerstin Ostermann (Institute for Employment Research) - Presenting Author
Dr Sebastian Lang ( Leibniz-Institut für Bildungsverläufe)
In relabeling the German unemployment benefits (“Arbeitslosengeld”) to basic income (“Bürgergeld”) in 2022, the public debate increasingly relies on the suggestion that being unemployed and feeling stigmatized are interrelated. Previous research largely confirms this suggestion by showing that both being unemployed and feeling stigmatized lead to psychological and labor market obstacles. However, quantitative research on the spatial and social context in which stigma-consciousness emerges and intensifies is still scarce. We combine rich survey data from the Panel Study Labour Market and Social Security (PASS) with fine-grained spatial context data from the Federal Employment Agency to test whether the local neighborhood affects the respondent’s feelings of being stigmatized as unemployed. In linking survey data to register data of the level of 1x1km residential grid cells, we directly measure the local prevalence of unemployment and argue that high levels of neighborhood unemployment can be considered as deviant behavior from a general employment norm. Following labelling theory and the social contagion approach, such deviant behavior should lead to decreasing levels of stigma-consciousness of the unemployed. Contradicting these predictions, random intercept models show that the neighborhood’s unemployment positively affects the individual’s stigma-consciousness as soon as it exceeds a total share of 30% unemployed. Additional analyses reveal that the neighborhood is not the spatial level of norm-setting. However, neighborhoods act as norm-enforcement environments which are strongest for neighborhoods comprising high as well as low-income residents. Hence, the presence of stigmatized and stigmatizing individuals is necessary for effective norm-enforcement. Taken together, the study comes with methodological and theoretical advances. We not only highlight different spatial levels of local norm setting and enforcement, but also provide a case study for how to combine survey data with highly reliable and fine-grained geospatial context data.
Dr Andreas Hartung (RPTU Kaiserslautern-Landau) - Presenting Author
Mr Benjamin Stefan (RPTU Kaiserslautern-Landau)
Mr Jonas Breihof (RPTU Kaiserslautern-Landau)
Mr Nils Hausbrandt (RPTU Kaiserslautern-Landau)
Mr Julian Steinbacher (RPTU Kaiserslautern-Landau)
Professor Stefan Ruzika (RPTU Kaiserslautern-Landau)
In our research as part of the “Ageing Smart” project, we combined indirectly georeferenced survey data from seven pilot municipalities in Germany with walking distances from a self-developed spatial accessibility model. This combined dataset was used to assess how older residents rate their local access to shopping facilities. The results showed significant differences between municipalities of different structural types, suggesting a higher tolerance of distance in non-urban areas.
We hypothesise that residents adjust their distance expectations according to the local conditions, including the density of services and transport situation, as well as to their resulting individual mobility patterns. We aim to correct for this by including multimodal route planning in our analyses. This approach generates routes not only for different modes of transport, but also for their most optimal combination. We link route alternatives to the survey data according to the individually reported mobility behaviour.
The objectives of this study are
- to optimise the prediction of satisfaction with local infrastructure with respect to individual mobility preferences and
- to clarify the differences between municipality types in terms of residents’ mobility behaviour and their according perceptions of (walking) distances.
The concept of the “city of short distances” is becoming one of the most dominant in urban and spatial planning. It postulates that urban spaces should be designed in such a way that most of the daily needs of residents can be met within a short walking distance. This may not be applicable in non-urban areas, not only in terms of feasibility, but also in terms of individual mobility expectations, which reflect, among other things, retrospective housing choices. With our results, we want to contribute to this discussion. In addition to this substantive contribution, we provide an interesting example of the promising combination of survey data with high quality spatial
Dr Annette Scherpenzeel (Nivel) - Presenting Author
Professor Michel Dückers (Nivel)
Concerns about the health impacts of living near industrial areas in the Netherlands have been rising, with studies highlighting elevated risks of chronic and acute health conditions in affected populations. This research builds on previous findings, such as the increased prevalence of lung cancer, COPD, and diabetes in industrial zones, particularly around the IJmond region and Tata Steel facilities, where emissions of heavy metals and carcinogens exceed expected levels.
Using the Dutch Health Monitor (GEMON) survey data, this study comprehensively assesses health disparities among residents living near industrial sites compared to the general population. GEMON provides harmonized and longitudinal health data from thousands of adults across the Netherlands, covering demographics, lifestyle factors, and self-reported health outcomes, including non-specific complaints, mental health issues, and chronic conditions. By linking GEMON data with income information available at Statistics Netherlands and environmental exposure data (e.g., air pollution and proximity to industrial sites), the study investigates the following aspects:
1. The differences in acute and chronic health outcomes between residents near industrial zones and the broader population, accounting for factors like age, gender, lifestyle, and household income.
2. Deviations in health risks in the IJmond region from national trends, given its previously documented elevated health risks.
3. The consistency of these health disparities over time using data from 2016, 2020, and 2022.
This quantitative study involves the integration of GEMON responses with geocoded exposure metrics and socioeconomic data. The findings address gaps in understanding the full spectrum of health impacts, particularly in vulnerable populations living near industrial areas.
Dr Dennis Abel (GESIS - Leibniz Institute for the Social Sciences) - Presenting Author
Dr Stefan Jünger (GESIS - Leibniz Institute for the Social Sciences)
A growing interest in economics and the social sciences in Earth observation (EO) data has led to a broad spectrum of publications in recent years. They range from studying environmental attitudes and behavior, economic development, conflicts and causes of flight, and electoral behavior. However, social science researchers also face many obstacles in applying and using these data, resulting from 1) a lack of technical expertise, 2) a lack of knowledge of data sources and how to access them, 3) unfamiliarity with complex data formats, such as high-resolution, longitudinal raster datacubes, and 4) lack of expertise in integrating the data into existing social science datasets. Despite the increased interest in the data, for the majority of researchers in the social sciences, EO data represents a black box after all. In this session, we present our new project “EarthLinks” which aims to close the gap and create an automated interface to EO data and complementary resources for social science research. The project's goal is creating an open-source tool to link time- and space-sensitive social science datasets with data from Earth observation programs based on a Shiny App in R. The project advances the automatization of these data integration processes between social science data and EO data based on an open-source, user-friendly tool that does not require users' programming skills. The EarthLinks workflow will be exemplified on the basis of a research project which we currently conduct on the effects of flooding exposure on climate change opinion.
Mr Jens Carstens (Sciences Po Paris) - Presenting Author
Ms Anne-Kathrin Stroppe (GESIS - Leibniz Institute for the Social Sciences)
Economic deprivation at the local level shapes political attitudes and behaviour, especially in recent elections across Western democracies. The prevailing assumption is that objective economic conditions act as information cues, shaping citizens’ subjective perceptions of local economic conditions. However, subjective economic perceptions often diverge from objective indicators. In this article, we examine the objective-subjective nexus by investigating the factors influencing citizens’ perceptions of a place’s economic deprivation. Specifically, we assess the role of absolute economic conditions, relative conditions across space, relative conditions over time, and comparisons to the most salient out-group, the economy in the capital. Analyzing five waves of pooled cross-sectional survey data from the British Election Study (2017-2022), linked to median house prices at the Middle Super Output Area (MSOA) and the Local Authority District (LAD) level as proxies for local and regional prosperity, our results reveal that while absolute local house prices strongly influence perceptions of a place’s economy, citizens also draw on relative comparisons. However, both spatially, comparing their locality to London, and temporally, assessing its local economic trajectory over time, effect strengths vary. These findings highlight the multidimensional nature of economic evaluations, underscoring the interplay between objective economic conditions, subjective perceptions, and salient reference points.