Measuring Social Contexts with Survey Data and Beyond – Applications and Methodological Challenges |
|
| Coordinator 1 | Dr Dominik Becker (University of Tuebingen) |
| Coordinator 2 | Professor Steffen Hillmert (University of Tuebingen) |
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.