PLoS By Category | Recent PLoS Articles

Public Health and Epidemiology - Science Policy

Presenting Survey Items One at a Time Compared to All at Once Decreases Missing Data without Sacrificing Validity in Research with Internet Volunteers
Published: Thursday, May 17, 2012
Author: Brian A. Nosek et al.

by Brian A. Nosek, N. Sriram, Emily Umansky

In two large web-based studies, across five distinct criteria, presenting survey items one-at-a-time was psychometrically either the same or better than presenting survey items all-at-once on a single web page to volunteer participants. In the one-at-a-time format, participants were no more likely to drop-out of the study (Criterion 1), and were much more likely to provide answers for the survey items (Criterion 2). Rehabilitating participants who otherwise would not have provided survey responses with the one-at-a-time format did not damage internal consistency of the measures (Criterion 3) nor did it negatively affect criterion validity (Criterion 4). Finally, the one-at-a-time format was more efficient with participants completing it more quickly than the all-at-once format (Criterion 5). In short, the one-at-a-time format results in less missing data with a shorter presentation time, and ultimately more power to detect relations among variables.