SRO develops and implements a range of sample design and selection protocols from multi-stage area probability samples to list-based and respondent-driven techniques. SRO services also include the design and construction of sample weights, non-response adjustment, and imputation.
Sample Design and Implementation
The present-day work of SRO builds on the foundation laid by Leslie Kish and his SRC colleagues who pioneered survey sampling techniques in the 20th century. SRO continues its prominence today through its involvement in sample selection for some of the nation’s most widely-used social science and public health data resources. Our expertise in sample design and selection is vast, including multi-stage area probability samples of the U.S. population, oversampling to study population subgroups, list-based samples of campus communities and professional society members, school-based samples of teacher and student populations, unit-based samples of military personnel, and respondent-driven sampling techniques. SRO sampling statisticians also support the implementation of methodological experiments, intervention studies, and randomized controlled trials to ensure accurate implementation in data collection.
Because samples do not perfectly match the characteristics of a study population, SRO will design and develop sample weights specific to your project needs. Sample weighting helps to improve the accuracy of survey results by adjusting for differences between the sample and the target population. This process can help reduce bias and improve the reliability and validity of survey findings. SRO will design and develop base weights, post-stratification weights, and non-response adjustment.
Post-survey Weighting
SRO staff are trained in the most current post-processing, weighting and imputation methods, and they boast extensive experience applying these methods in the creation of non-response adjustments, sampling weights and imputed values. SRO statisticians use non-response adjustment to correct for potential response bias introduced by under-representation of population sub groups. This process ensures that the final sample of respondents reflects the true distribution of the population. Post-stratification weights based on known population demographics ensure that the sample aligns with the population distribution.
Researchers who wish to consult with SRO on sampling or post-survey adjustment should Submit a Survey Request Form.