On Spillover Effects of Social Image Concerns

How to cite

Oslislo, Christoph; Suchy, Rafael; Grabe, Leonhard. 2022. "On Spillover Effects of Social Image Concerns." AEA RCT Registry. AEARCTR-0009554. https://doi.org/10.1257/rct.9554-1.1

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

Christoph Oslislo* , Rafael Suchy , Leonhard Grabe

* Primary investigator

Registry ID
AEARCTR-0009554
DOI
https://doi.org/10.1257/rct.9554-1.1
First registered
Last updated
Status
Completed
Trial dates
2022-06-08 – 2022-10-31
Countries
Germany ()
Registry record
http://www.socialscienceregistry.org/trials/9554

Abstract

A feeling of being observed can lead individuals to behave more prosocially. Based on a static comparison of decision-making in public versus in private, this finding has been shown to be very robust across various domains of ethical decision-making. Social image concerns - often claimed as the underlying reason - are regularly exploited by charities, governments and firms in practice. However, many real-life applications are characterized by repeated decision-making and there are many contexts in which it is not possible (or too costly) to always ensure observability of actions. In this project, we explore dynamics of prosocial decision-making with a focus on the role of observability by means of an online lab experiment. In particular, we investigate whether observability of initial actions in a specific prosocial domain affects subsequent private decision-making in the same domain.

Intervention

We conduct an online lab experiment and analyze behavior in a variant of the dictator game. Participants are repeatedly confronted with a decision situation in which they divide money between themselves and a charity. Building on research on moral balancing, our analysis focuses on the effects of a variation of circumstances of initial behavior on subsequent behavior in the same domain. In this spirit, we design a setting in which we exogenously vary the observability of initial behavior in a controlled manner while we hold the circumstances of private subsequent behavior constant across treatments. This setting allows us to examine our main research questions: Does observability of behavior lead to different subsequent private behavior and what drives positive and negative spillover effects, i.e. moral consistency and moral licensing, respectively?

Experimental design

The experiment consists of three parts: A pre-experimental questionnaire, a behavior experiment, and a post-experimental questionnaire. The experimental design of the main part of the experiment is a variant of a repeated dictator game over T = 4 rounds. The specific design allows us to study the effects of observability of initial behavior (rounds 1 and 2) on subsequent behavior (rounds 3 and 4) in the same domain of ethical decision-making. Each round consists of a simple real-effort task (RET) followed by a donation decision. In the real-effort task, participants earn an endowment of 7 EUR which they are asked to split between themselves and a charity (World Food Programme). Between treatment and control group, we vary observability of the RET and the donation decision of rounds 1 and 2: Participants in the treatment group are observed by another person via life webcam during rounds 1 and 2 and see themselves in rounds 3 and 4. Participants in the control group see themselves over the whole course of the main part of the experiment, i.e. rounds 1 to 4.

Design details

Randomization method
We randomly assign participants to either the control or the treatment group. Randomization is done by the experimenter who randomly assigns two different weblinks to participants based on the random order in which participants entered a zoom call right before the start of each session (serving the purpose of an attendance and camera check).
Randomization unit
individual level
Clusters
no clusters, randomization on individual level
Observations
320 participants from the CLER participant pool
Arms
160 participants in each of the two treatment arms
Primary outcomes
Our main variables of interest are donation decisions across four rounds of a dictator game. Key outcome variables will be within-subject differences in decisions across rounds and between-subject differences in private round 3 (between Control and Treatment).
Secondary outcomes
As secondary outcome variables, we measure several individual-specific characteristics: - Demographics (age, gender, occupation, donation patterns) - Intrinsic altruism - Preferences for consistency - Situational self-awareness and self-perception after round 1 and after round 3 - Perceived social pressure after round 1 and after round 3 - Cognitive Reflection Test - Personal views on Moral Licensing

Cited by

No citation data available for this registration yet.

Keywords