How we reach respondents
Survata reaches respondents on publisher sites around the web. We partner with publishers to let their visitors take a Survata survey to unlock the publisher’s premium content (e.g., premium articles, ebooks, videos, etc.). An example of Survata survey unlocking premium content on a publisher site can be found here. All of our publishers are required to have our surveys unlock content; none of our surveys appear in banner ads, in “pop-ups,” or in “sidebars” on publisher sites. Publishers must be approved to join our network (if you are a publisher, you can apply here).
Survata does not maintain an online panel or mailing list. We do not reach respondents by telephone or in-person. All of our surveys are done anonymously.
How we ensure accuracy
Our primary focus of every survey we run is to deliver accurate results. This means the results of a subpopulation we sample should reflect the expected results from surveying an entire population. The components of accurate results are ensuring respondents answer honestly (the “veracity” of the answers), ensuring we reach a representative sample of a subpopulation (the “bias” of a sample), and ensuring we reach an adequate number of respondents to be statistically significant (the “margin of error” of the results).
Veracity of answersSurvata employs multiple approaches to ensure we reach quality respondents answering honestly. First, we screen every publisher who applies to join our network to verify they attract high quality, attentive visitors to their site. Only a small fraction of publishers who apply to join our network are approved to join. Second, we have developed technology to detect when respondents are not answering honestly. Our Online Lie Detector™ technology tracks a respondent’s response time, response pattern, and other metadata to determine which responses should be discarded. A discarded response is not included in a client’s results, and the client is not charged for it. Third, we constantly test the health of our network with test questions and “trap” questions that identify respondents who are not answering honestly.
Bias of sampleSurvata is careful to minimize bias in a survey’s sample, meaning we avoid over-indexing or under-indexing on certain traits of respondents. For example, running a survey with 90% of respondents being female might lead to different results than a survey with 50% of respondents being female. For nationally representative surveys (such as our “Standard Read” and “Deep Read” packages) we ensure we are U.S. census-representative for gender, age group, and geography. We can also look at potential biases for other traits based on a client’s specific needs.
Margin of error of resultsSurvata can deliver statistically significant results to our clients, meaning the results are outside of the margin or error and thus likely not due to chance alone. Some clients who utilize our “Quick Read” package are comfortable with a “directionally correct” answer with a higher margin of error. Other clients aim for a lower margin of error (usually below 4%). Survata can advise clients what margin of error is appropriate for their survey.
Like any sample provider, we have known biases in the respondents we reach. We only run surveys on websites, so we do not reach respondents without Internet access, or respondents with Internet access who never never visit websites. According to the latest data from the U.S. Census Bureau, 84.1% of individuals above age 3 have some form of Internet access. Also, due to U.S. regulations, we do not survey respondents under the age of 13. Therefore all of our survey results are biased towards individuals aged 13 and above.