An experimentally tested and externally-validated approach to calibrating non-probability and probability data that is proven to reduce bias and improve data quality.
Reasons to Consider SSRS Hybrid Samples
Research consistently finds that probability-based samples remain the “gold standard” for projecting estimates to a larger population but can be cost-prohibitive for many clients.
Nonprobability samples offer a much lower cost-per-complete but yield biased population estimates even if weighted on demographics because self-selection into online samples is driven by characteristics, attitudes, and behaviors that are not captured by traditional weighting demographics.
SSRS has developed an approach that utilizes data science to create a fully-customized and affordable solution to each unique application of hybridized data. This allows us to take advantage of both the low cost of non-probability sampling (to obtain larger sample sizes) and the statistical rigor of probability sampling (to reduce the risk that estimates will be biased).
SSRS Hybrid Calibration Solution Offerings
Unique Stepwise Methodology
Our unique stepwise calibration methodology for choosing the calibration model that is most effective at reducing selection bias and total error in key study outcomes. This experimentally validated procedure adapts a guided-search algorithm that maximizes bias reduction in a cost-effective way.
Expansive Bank of Calibration Items
An expansive bank of topic-customizable non-demographic calibration items validated through rigorous experimentation. These items offer an alternative to “one-size-fits-all” weighting by tailoring calibration items to the study topic, improving the effectiveness of calibration and controlling selection bias in study outcomes.
A full-service research organization, our highly experienced SSRS Methods Analytics and Data Science (MADS) group conceptualizes and manages survey data collection from design to dissemination. They have extensive expertise in methodological experimentation, sampling, weighting, data collection planning and monitoring, data editing and imputation, disclosure prevention, documentation, reporting, and analysis.
How can we help you design and execute your research?