A Suite of Innovative Tools for Political and Election Polling
- SSRS Utilizes Mixed Methods Tailored to Research Needs
- Conducting “gold standard” mixed-mode address-based sample (ABS) designs to obtain population benchmarks for non-demographic attitudinal measures, which help measure and reduce nonresponse biases in faster-turnaround polls.
- Using the SSRS Opinion Panel and the SSRS Text Message Panel for rapid-turnaround research.
- Applying our Encipher Hybrid calibration approach to increase available sample sizes, especially at the state level, by blending probability and nonprobability samples
- SSRS Links a Variety of Data Sources to Improve Survey Data Quality
- Appending external data (voter and commercial files, small area demographics, and precinct-level election results) to detect and correct for partisan nonresponse.
- Applying machine learning models to predict the party affiliation of persons at an address for stratification and weighting.
- Assigning likely voter scores based on vote history and other externally available data.
- Validating self-reports of past voting behavior.
- SSRS Applies an Advanced Use of Weighting Adjustments
- Weighting to new non-demographic benchmarks (e.g., recalled vote and modeled party identification within small geographic areas) to correct for differential response rates by party and other non-demographic drivers of polling error
