Data Linking - Linking survey data with geospatial, social media, and sensor data (Version 1.0)

Abstract

Survey data are still the most commonly used type of data in the quantitative social sciences. However, as not everything that is of interest to social scientists can be measured via surveys, and the self-report data they provide have certain limitations, such as recollection or social desirability bias, researchers have increasingly used other types of data that are not specifically created for research. These data are often called "found data" or "non-designed data" and encompass a variety of different data types. Naturally, these data have their own sets of limitations. One way of combining the unique strengths of survey data and these other data types and dealing with some of their respective limitations is to link them. This guideline first describes why linking survey data with other types of data can be useful for researchers. After that, it focuses on the linking of survey data with three types of data that are becoming increasingly popular in the social sciences: geospatial data, social media data, and sensor data. Following a discussion of the advantages and challenges associated with linking survey data with these types of data, the guideline concludes by comparing their similarities, presenting some general recommendations regarding linking surveys with other types of (found/non-designed) data, and providing an outlook on current developments in survey research with regard to data linking.