Testing Experimental Robustness across Survey Invitation Strategies

How to cite

Hansen, Paw. 2021. "Testing Experimental Robustness across Survey Invitation Strategies." AEA RCT Registry. AEARCTR-0007408. https://doi.org/10.1257/rct.7408-1.0

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

Paw Hansen*

* Primary investigator

Registry ID
AEARCTR-0007408
DOI
https://doi.org/10.1257/rct.7408-1.0
First registered
Last updated
Status
In development
Trial dates
2021-04-06 – 2021-04-27
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/7408

Abstract

We randomly assign three distinct invitations to participate in a survey. Each invitation is hypothesized to trigger a particular form of motivation with the potential respondents. The overall aim is to test (a) whether these subtle differences in the invitation to participate (a) affect response rates, and (b) shape the sample due to self-selection and. Moreover, we embed a separate survey experiment in the survey and test whether the three different invitations lead to treatment heterogeneity.

Intervention

We randomly assign participants to receive one of three distinct invitations to participate in a survey. Each invitation appeals to a specific form of motivation. Thus, the intervention consists of subtle text cues in the invitations.

Experimental design

We randomly assign participants to receive one of three distinct invitations to participate in a survey. Each invitation appeals to a specific form of motivation. Thus, the intervention consists of subtle text cues in the invitations. Within the survey, we embed a survey experiment. Our goal is to examine differences between the three groups in terms of (a) response rates (b) levels of motivation, and (c) heterogeneity in how respondents react to the embedded survey experiment. The embedded survey experiment is preregistered separately.

Design details

Randomization method
Randomization is done with computer.
Randomization unit
We conduct randomization at the individual level.
Clusters
Approx. 9,000 public employees.
Observations
Approx. 9,000 public employees.
Arms
Approx. 3,000 public employees per treatment arm.
Primary outcomes
We have three outcomes of interest. (1) Response rates across the three groups, (2) respondents' levels of three types of motivation (Public Service Motivation, user orientation, and self-efficacy), and (3) heterogeneity in how respondents reply to a survey experiment embedded in the issued survey

Cited by

No citation data available for this registration yet.

Keywords