Evaluating Online Nonprobability Surveys
Online nonprobability surveys are fast, cheap, and increasingly popular. We compared nine samples and found that accuracy varied substantially.
Many Americans say they voted, but did they?
One-in-six (16%) of those who say they “definitely voted” in the 2014 midterm election have no record of voting in commercially available national voter files.
Data Quality Deserves an Honest Discussion
While the possibility of falsified data is an important consideration in survey research, a new tool to detect it fails to perform as advertised.
Evaluating a New Proposal for Detecting Data Falsification in Surveys
The claim by Kuriakose and Robbins (2015) that there is widespread falsification in international surveys is clearly concerning. However, an extensive investigation conducted by Pew Research Center finds the claim is not well supported.
Can Likely Voter Models Be Improved?
High-profile polling failures in recent elections have drawn attention to the challenges in using surveys to predict outcomes. Our study examines various methods of determining who is a likely voter.
Pew Research Center will call 75% cellphones for surveys in 2016
We’re making this change to ensure our survey samples properly represent the now roughly half (47%) of U.S. adults who only have a cellphone.
Advances in Telephone Survey Sampling
Telephone surveys face numerous challenges, but some positive developments have emerged, principally with respect to sampling.
The Unique Challenges of Surveying U.S. Latinos
Surveying Hispanics is complicated for many reasons – language barriers, sampling issues and cultural differences – that are the subject of a growing field of inquiry.
Who’s left out in a Web-only survey and how it affects results
We surveyed non-Web panel members by mail and assessed how much, if at all, their non-participation would affect the outcome in a poll conducted exclusively online.
Coverage Error in Internet Surveys
With 89% of U.S. adults online, survey research is rapidly moving to the Web. But 89% is not 100%, and surveys that include only those who use the internet run the risk of producing biased results.
