Surveys meet social media
Blending traditional and non-traditional data sources
Research Goals
Using SSRS’s probability-based data collection infrastructure for reliable estimates, one goal of this project is to collect a large enough sample of Twitter users to link micro-level survey data to social media data and improve our understanding of the measurement properties of social media.
The project team aims to reliably study the following sub-groups:
- Have Twitter, gave consent for Twitter handle
- Have Twitter, did not give consent for Twitter handle
- Does not have Twitter
Multiple waves of surveys will be deployed to understand what constructs can or cannot be mapped to topics of interest to social science researchers. Using the initial survey from the full panel, the research team published a methods brief estimating Twitter usage by state.
Suggested Citation: Jackson, M. McKinstry, J., McPhee, C., Raghunathan, T., Singh, L., Traugott, M., Turakhia, C. & Wycoff, N. (2021). Understanding who uses Twitter: State level estimates of those on Twitter. MOSAIC Methods Brief: November 2021. Measuring Online Social Attitudes and Information Collaborative.
Linking Surveys to Twitter Posts
With the Linkage, we will be able to:
- Understand who is producing Twitter posts not only who is a user of Twitter.
- Gain understanding of potential population inference
- Validate user-level predictions from Twitter posts by using survey data as “ground truth”.
- Understand what constructs can or cannot be mapped to topics of interest to social science researchers.
This first dataset providing preliminary state-level estimates of Twitter usage is available through the MDI portals.
