Social Norms and Online Survey Participation

How to cite

Brade, Raphael; Jäckle, Robert. 2022. "Social Norms and Online Survey Participation." AEA RCT Registry. AEARCTR-0009303. https://doi.org/10.1257/rct.9303-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Raphael Brade* , Robert Jäckle

* Primary investigator

Registry ID
AEARCTR-0009303
DOI
https://doi.org/10.1257/rct.9303-1.0
First registered
Last updated
Status
On going
Trial dates
2022-03-15 – 2023-10-31
Countries
Private
JEL codes
C83, C93, D91, H41
Registry record
http://www.socialscienceregistry.org/trials/9303

Abstract

Researchers often rely on data from voluntary (online) surveys, for instance, to investigate effects on outcomes that are not available in administrative data (e.g., well-being and non-cognitive skills) or to study potential mechanisms that drive effects (e.g., behavioral changes may occur due to an update of individuals’ beliefs). However, survey response rates are often low, which reduces the statistical power of subsequent analyses. In addition, survey participation may be selective, which can affect nonresponse bias. Using a large sample of university students, we investigate if social norms can be used to increase participation rates and to reduce the nonresponse bias in a voluntary online survey.

Intervention

The intervention will be implemented as part of the invitations to two online surveys that are conducted in the summer semester 2022 with students of all bachelor’s programs at a university of applied sciences studying towards 27 different degrees. In all experimental groups, students will receive an email invitation and four email reminders asking for their participation in the surveys.

Experimental design

Students were randomly allocated to a control and seven treatment groups.

Design details

Randomization method
Blocked randomization using Stata
Randomization unit
Student-level
Clusters
9,276 students
Observations
9,276 students
Arms
Group 1: 1,161 Group 2: 1,162 Group 3: 1,159 Group 4: 1,163 Group 5: 1,153 Group 6: 1,162 Group 7: 1,156 Group 8: 1,160
Primary outcomes
Survey participation
Secondary outcomes
An increase in survey participation can, but need not, increase survey quality. To assess how survey quality is affected by the treatments and study potential mechanisms, we will also investigate effects along the following dimensions: 1. Nonresponse bias: Nonresponse bias will be assessed by approaches such as analyzing differences in observable administrative background variables between respondents and nonrespondents (such as age, sex, features of the university entrance qualification, and prior academic performance). 2. Quality of survey response: We will further approximate the quality of the survey responses using variables based on dimensions such as answering speed, item-non-response, survey drop-out, and item-non-differentiation. 3. Beliefs: Finally, to study potential mechanisms we will investigate effects on students’ beliefs about the participation rate among all students and the confidence in those beliefs (asked in the surveys).

Cited by

No citation data available for this registration yet.

Keywords