ESRA 2019 Programme at a Glance
Measuring Social Contexts with Survey Data and Beyond – Applications and Methodological Challenges
|Session Organisers|| Dr Dominik Becker (University of Tuebingen)
Professor Steffen Hillmert (University of Tuebingen)
|Time||Friday 19th July, 11:00 - 12:30|
A growing body of social research has been concerned with analysing contextual effects. Such analyses entail several substantial and technical challenges. This session invites submissions from all fields of application that deal with the challenges of defining relevant contexts, generating adequate contextual data, and matching contextual with survey data to measure corresponding contextual effects.
Researchers today have access to a variety of public and commercial surveys, administrative data, and other forms of data such as sensor data, social media or transaction data. However, the availability of relevant contextual data and access to it still constitute major problems. Moreover, having substantive data at hand does not necessarily imply that the construct of interest is measured with sufficient precision. Often, social scientists are interested in the effects of constructs that are located on a higher level of aggregation, while the indicators provided in the data have been measured on a lower level.
In the most straightforward setting, lower-level indicators can map both lower-level and higher-level constructs. For instance, students' socioeconomic status is an individual-level construct that is measured by individual-level indicators. The resulting scale can be aggregated to measure socioeconomic composition on the school level. In a less straightforward setting, the construct makes theoretical sense only at a higher level. Yet, due to either methodological interest or data availability, survey researchers will often use lower-level indicators to measure the higher-level construct. For instance, teaching quality can be measured by students' ratings, but these ratings only represent a meaningful construct on the classroom level. Consequently, a corresponding scale should be generated on this higher level of aggregation.
We welcome contributions that address the topic of measuring social context conditions from either substantive, methodological, or theoretical perspectives. Substantive contributions would typically present practical applications from ongoing research, which make use of surveys or other forms of disaggregated data. These might entail classification techniques drawing on multiple lower-level indicators for the measurement of social context conditions on the level of schools, neighbourhoods, countries, occupations, etc. Methodological contributions could compare results of different aggregation or classification techniques, illustrate potential sources of bias, or highlight how data from different sources (e.g., survey data, administrative or social media data) can be fruitfully combined to measure relevant aspects of social contexts. Theoretical contributions could outline differences in the causal pathways between effects of aggregated and generic characteristics of social contexts.
Keywords: Contextual effects, aggregation, classification, data linkage
Survey Measures of Social Capital: How Well Do Self-Reported Indicators Measure Social Capital?
Professor Stefano Bartolini (University of Siena)
Professor Paola Bordandini (University of Bologna)
Professor Roberto Cartocci (University of Bologna)
Dr Francesco Sarracino (STATEC) - Presenting Author
Despite the importance and wide application of social capital in social sciences, the validity of its measures hinges on experimental and anecdotal evidence, and on correlations among survey variables. We perform a large scale test of the validity of survey measures of social capital, and in particular of trust in others, by comparing survey and objective measures across countries, regions and over time. Our evidence suggests that survey measures reliably reflect objective measures of social capital.
Linking Administrative School Data with Survey Data to Analyze Contextual Effects of Learning Environments
Dr Uta Landrock (LIfBi - Leibniz Institute for Educational Trajectories) - Presenting Author
Dr Sabine Zinn (LIfBi - Leibniz Institute for Educational Trajectories)
This presentation explores the feasibility of accessing and linking administrative school data with the survey data of the National Education Panel Study (NEPS). The NEPS data were collected at the individual level of students and additionally at the contextual level of teachers and school principals. Participation is voluntary, therefore the information is incomplete: Information is not available for all entities (unit non-response) and not all relevant information can be collected in surveys, particularly at the contextual level. Administrative data at the school level, in contrast, are complete as the data are collected by standard.
We show the potential of linking administrative data with NEPS data and the challenges with regard to data availability, accessibility and harmonization applying the example of empirical educational research: Current findings indicate that school achievements (e.g. grades) depend not only on the students themselves, but also on their learning environments. Information on these contextual factors is not fully available in the NEPS data, but has been collected as administrative data by the statistical offices of the federal states. In Germany there are 14 statistical offices which collect, process and make data available. Data accessibility concerns data protection issues, legal restrictions, complex administrative application and negotiation procedures as well as linking data from up to 14 statistical offices which are not comparable in all variables.
Taking North Rhine-Westphalia as an example, we show that linking administrative and survey data increases analytic potential. Especially, at the contextual level of schools there is added value in administrative data. Additional information is, for example, characteristics and degrees of school leavers or the sociodemographic composition of teachers at a school.
Precarity of Job Entry Histories, Its Subjective Perception and the Vocational Field of Qualification
Dr Ralf Dorau (German Federal Institute of Vocational Education and Training) - Presenting Author
The passage from education into work is an important transition in life. Therefore, we analyze the occupational integration of transition sequences into work during the first three years after finishing the first degree in vocational training or university studies. To evaluate histories in longitudinal studies we have developed two indices, one for integration and one for disaffiliation. Both include parameters for the total time of statuses, assigned to a particular zone of occupational integration in the observation period, the duration of continuous statuses and the number of employment interruptions. This enables us to classify job histories with regard to occupational integration and compare them between cohorts or even different datasets. For young skilled employees in Germany there are around 30 % precarious and about 5 % disaffiliated in the late 2000s.
However, does the subjective evaluation of the three-year-sequence as a more or less successful transition to work by the respondents match the ‘objective’ criteria provided by the indices of integration or disaffiliation? Alternatively, does the subjective perception of the occupational integration at the career entry depend to a greater extent on other variables like a particular vocational education, school leaving certificates, region, gender or parent’s education?
Multilevel analysis enables us to analyze the probability of positive or negative subjective perception of the respondents’ transition sequences with the vocational field of a particular qualification on a structural level. Additionally we analyze context effects, for instance unemployment rates, and additional individual effects like parents’ school leaving certificates or gender. Our research is based on a survey in the framework of BIBB/BAuA Labour Force Survey 2012 with job entry processes of graduates 2006 to 2008 (N=4.446).
Using Occupation-Level Working Conditions to Compare Effects of Vertical and Horizontal Social Mobility on Individual Subjective Well-Being
Dr Dominik Becker (University of Tuebingen) - Presenting Author
Professor Steffen Hillmert (University of Tuebingen)
This contribution uses employees’ reports of job tasks to measure working conditions at the level of occupations. In a second step, this measure of working conditions is used to study effects of social mobility on subjective well-being (SWB) at the level of individuals. Substantively, we advance conventional arguments on social mobility effects by the assumption that not only upward and downward, but also horizontal social movements in terms of changes in working conditions may cause psychological distress and thereby lower individual SWB. Moreover, horizontal mobility effects on SWB might be stronger for intermediate classes, and vertical and horizontal mobility effects might reinforce one another.
Empirically, we use the BIBB/BAuA Labour Force Survey 2011/2012 to identify relevant job task dimensions of ISCO-08-coded occupations. By means of multilevel confirmatory factor analysis, we compute the occupation-level latent factors of five job-task dimensions (non-routine analytic, non-routine interactive, routine cognitive, non-routine manual and routine-manual job tasks). In a second step, we merge this information to the 2012 and 2013 waves of the German Socio-Economic Panel in order to analyze intragenerational occupational mobility effects on individual SWB. We compare two ways of identifying the horizontal dimension of job task mobility: first, we control effects of job tasks for established indicators of vertical stratification such as occupational prestige and income. Second, we aim to isolate the vertical and the horizontal dimension of working conditions as latent factors directly from the measurement model. First analyses reveal a considerable variation of job tasks over and above indicators of vertical stratification. This allows for a meaningful analysis of both vertical and horizontal social mobility effects on SWB.