To accurately measure public opinion, pollsters need samples that are representative of a larger population. Researchers often do this by drawing random samples from lists of all telephone numbers or mailing addresses, but it’s much more challenging when they want to conduct surveys online. That’s because there’s no master list of all email addresses for the entire population.
While some organizations (including Pew Research Center) use traditional offline methods like telephone or postal mail to collect email addresses from a random sample of the public, many online surveys are conducted using “nonprobability” or “opt-in” samples. These polls are generally easier and cheaper for polling organizations to conduct, but questions about their accuracy remain.
In this video explainer – the latest in our Methods 101 series – we explore some of the features of nonprobability surveys and how they differ from traditional probability-based polls. To find out more about these kinds of polls, you can read two evaluations we conducted, one in 2016 (summarized in this Q&A) and another in 2018.[callout]
Watch our other Methods 101 videos:
How can a survey of 1,000 people tell you what the whole U.S. thinks?
How do you write survey questions that accurately measure public opinion?
Phone vs. online surveys: Why do respondents’ answers sometimes differ by mode?
How is polling done around the world?
What is machine learning, and how does it work?[/callout]