Content Matters: The Effects of Commitment Requests on Truth-Telling, UPDATE

How to cite

Rincke, Johannes; Glogowsky, Ulrich; Cagala, Tobias; Schudy, Simeon. 2022. "Content Matters: The Effects of Commitment Requests on Truth-Telling, UPDATE." AEA RCT Registry. AEARCTR-0009683. https://doi.org/10.1257/rct.9683-2.0

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

Johannes Rincke* , Ulrich Glogowsky , Tobias Cagala , Simeon Schudy

* Primary investigator

Registry ID
AEARCTR-0009683
DOI
https://doi.org/10.1257/rct.9683-2.0
First registered
Last updated
Status
Completed
Trial dates
2022-07-13 – 2022-11-30
Countries
Germany ()
Registry record
http://www.socialscienceregistry.org/trials/9683

Abstract

This trial is an update of Cagala, Tobias, Ulrich Glogowsky and Johannes Rincke. 2020. "Content Matters: The Effects of Commitment Requests on Truth-Telling", https://doi.org/10.1257/rct.6700. We plan to collect additional data on the cheating game in a laboratory experiment and an online survey on psychological reactance.

The main changes relative to the cited trial are as follows: First, compared to the last run, the experimental data will be collected in the laboratory rather than online. This step allows us to test the robustness of our results across settings. Second, we will collect data in two laboratories (Munich and Konstanz). To facilitate data collection, we extended our team of principal investigators (Simeon Schudy has joined the team as a new PI). Third, we will reduce the number of treatments to maximize the statistical power. Specifically, we will implement the control group, the ethics treatment, and the sanction treatment (but not the neutral treatment). Fourth, we decided to change the ordering of the reactance survey and the experiment: The experimental data will be collected first, followed by the online survey on reactance. Hereby, we avoid any spillovers of the survey questions on behavior in the experiment.

Intervention

We plan to repeat the trial described in AEARCTR-0006700 with a few changes: First, compared to the last run, the experimental data will be collected in the laboratory rather than online. This step allows us to test the robustness of our results across settings. Second, we will collect data in two laboratories (Munich and Konstanz). To facilitate data collection, we extended our team of principal investigators (Simeon Schudy has joined the team as a new PI). Third, we will reduce the number of treatments to maximize the statistical power. Specifically, we will implement the control group, the ethics treatment, and the sanction treatment (but not the neutral treatment). Fourth, we decided to change the ordering of the reactance survey and the experiment: The experimental data will be collected first, followed by the online survey on reactance. Hereby, we avoid any spillovers of the survey questions on behavior in the experiment.

Experimental design

Parts: The trial consists of two parts: an experiment, and a survey on psychological reactance.

Timing: The experimental data on the cheating game will be collected first, followed by the online survey on psychological reactance.

Part I (laboratory experiment): Our primary goal is to study whether different commitment requests (ethics and sanction) do affect cheating behavior relative to the control group. Our design is as follows: Subjects play a standard cheating game. A random draw determines a number between 1 and 6. Subjects receive an additional payoff of 5 Euro if they report a 5, and no additional payoff if they report any other number. In this setting, we test if different commitment requests affect cheating behavior. Particularly, at the beginning of the experimental sessions, subjects in the treatment groups sign a no-cheating declaration (control group: no declaration, no signature). The content of the declaration varies between a version referring to the standards of ethical sound behavior (ETHICAL STANDARD) and a version with a threat that non-compliance will be sanctioned (SANCTION). To maximize our statistical power, we do not implement the NEUTRAL treatment.

Part II (reactance survey): Subjects will be invited to participate in the online survey about two weeks after the laboratory session. The survey elicits psychological reactance in the same way as in the original trial. The reactance questions are embedded in a Big-5 questionnaire with 15 questions. We will also consider the heterogeneity of the treatment effects with respect to the subjects' reactance score as a secondary analysis. By running the experimental sessions before eliciting reactance, we make sure that the cheating behavior in the experiment cannot be confounded by the reactance survey (a concern raised by one of the referees).

Data linking: The experimental data and the survey data will be linked via a random code. Subjects revive the code upon entering the laboratory on paper and will be asked to type it into a text field both in the laboratory and when participating in the online survey.

Laboratories: The experimental sessions will be conducted in Munich and Konstanz, and if recruitment efforts require in Nuremberg. We need several laboratories in order to collect the planned number of observations.

Data analysis: The data analysis will follow the guidelines set by the original trial. As secondary analysis, we will study heterogeneity of the treatment effect regarding psychological reactance.

Design details

Randomization method
Randomization done by a computer
Randomization unit
Participant
Clusters
540 participants
Observations
We plan to collect data on 450 subjects who have an incentive to misreport in our experiment. Due to the nature of the random draw (for each subject, there is a 1/6 chance that the subject will have no rational option to cheat), we aim to recruit in total 540 participants (540*5/6) = 450). Due to attrition, the sample size for subjects who also complete the reactance survey two weeks later will be somewhat smaller. We will recruit subjects in the laboratories in Munich, Konstanz, and (if necessary) Nuremberg, but given the current situation regarding the Covid-19 pandemic, we cannot guarantee that we will succeed regarding the planned sample size.
Arms
150 subjects by treatment arm (after excluding subjects who, due to their random draw, had no incentive to cheat).
Primary outcomes
Indicator for subjects who cheat (misreport a drawn number)

Cited by

No citation data available for this registration yet.

Keywords