Survey Methods

Endres, Kyle, D. Sunshine Hillygus, Matt DeBell, and Shanto Iyengar. 2023. "A Randomized Experiment Evaluating Survey Mode Effects for Video Interviewing" Political Science Research and Methods. 11(1): 144-159.  pdf

Rising costs and challenges of in-person interviewing have prompted major surveys to consider moving online and conducting live web-based video interviews. In this paper, we evaluate video mode effects using a two-wave experimental design in which respondents were randomized to either an interviewer-administered video or interviewer-administered in-person survey wave after completing a self-administered online survey wave. This design permits testing of both within- and between-subject differences across survey modes. Our findings suggest that video interviewing is more comparable to in-person interviewing than online interviewing across multiple measures of satisficing, social desirability, and respondent satisfaction.   

Panagopoulos, Costas, Kyle Endres, and Aaron Weinschenk. 2018. "Preelection Poll Accuracy and Bias in the 2016 U.S. General Elections." Journal of Elections, Public Opinion & Parties. 28(2): 157-172. pdf

This report examines accuracy and bias in national- and state-level preelection polls conducted during the 2016 U.S. general election cycle. Overall, national polls in 2016 were somewhat more accurate than in 2012, but statewide polls were less accurate. Patterns across the board suggest polls underestimated Republican support in the presidential, U.S. Senate and gubernatorial races. Nevertheless, these biases were generally statistically insignificant, suggesting significant bias in preelection polls was scarce in 2016. 

Endres, Kyle, Erin O. Heiden, Ki Park, Mary E. Losch, Kari K. Harland, and Anne L. Abbott. 2023. "Experimenting with QR Codes and Envelope Size in a Statewide Push-to-Web Survey." Journal of Survey Statistics & Methodology. 

Survey researchers are continually evaluating approaches to increase response rates, especially those that can be implemented with little or no costs. In this study, we experimentally evaluated whether or not including Quick Response (QR) codes in mailed recruitment materials for self-administered web surveys increased web survey participation. We also assessed whether mailing these materials in a non-standard envelope size (6x9 inch) yielded a higher response rate than invitations mailed in a standard, #10 envelope (4.125x9.5 inch). These experiments were embedded in a sequential mixed-mode (dual-frame phone & web) statewide survey. Including a QR code (in addition to a URL) significantly increased the response rate compared to invitations that only included a URL in our study. As expected, a consequence of including the QR code was an elevated number of completions on smartphones or tablets among households randomly assigned to the QR code condition. The use of a larger (6x9 inch) envelope did not affect the overall response rate but did significantly boost the response rate for the landline sample (envelopes addressed to “STATE resident”) while having little effect for the wireless sample (envelopes addressed by name).  This study suggests that incorporating both QR codes and larger (6x9 inch) envelopes in mail recruitment materials for web surveys are cost effective approaches to increase web participation.