Westat, the Pew Research Center, and SurveyMonkey are collaborating to experiment with non-probability sampling of persons using Internet samples. The goal is to explore methods and tools that can be used with these new technologies to provide useful data in an era when contacting survey respondents and then gaining their cooperation is ever more difficult.
“Through this collaboration, we hope to better understand the nature of the biases in self-selected samples and whether any statistical techniques can be used to improve the quality of estimates from non-probability samples from Internet panel samples,” noted Westat Vice President and Senior Statistician J. Michael Brick, Ph.D.
The research will examine the science behind the adjustments and weighting made to four different methods of sampling: non-probability Internet-based; address-based; and probability-based phone and Internet. In addition, the research will explore “fit for purpose” use cases to help identify which sampling methods are best suited for various research objectives.
Westat will be providing its expertise in innovative sampling methods and will examine what other analytic techniques can be used with non-probability methods to increase their utility. Westat conducted a mail, self-administered survey that is being used in the investigation.
The Pew Research Center will draw on its work with its American Trends Panel to provide an alternative source of data from both telephone and a probability-based Internet panel.
SurveyMonkey, the world’s most widely used online survey software, has a sizeable respondent pool and conducted a survey from its panel. SurveyMonkey is well suited to initial experimentation because of the diversity of people who take surveys on its platform.
For additional media coverage on this news, see this video link by Bob Lederer of the Research Business DAILY report.