The impact of observability on pro-social behavior

We will implement six experimental conditions and one control condition:
• Control condition: Participants will play a standard dictator game. There is no observer in this condition.
• Decision Treatment: The decision maker’s choice of the amount sent to an anonymous receiver is shared with an anonymous observer.
• Decision plus Feedback Treatment: Like the Decision treatment, but the observer can give feedback on the decision maker’s choice. Specifically, the observer can select either a “like” or a “dislike” button.
• Photo Treatment: Like the Decision treatment, but additionally the decision maker and the observer exchange photos after the decision has been made.
• Photo plus Feedback Treatment: Like the Photo treatment, but the observer can give feedback on the decision maker’s choice. Like in the Decision treatment, the observer can select either a “like” or a “dislike” button.
• Video Treatment: Like the Photo treatment, but instead of exchanging photos, decision maker and observer meet in a video call where they can see each other but cannot communicate verbally.
• Video plus Feedback Treatment: Like the Video treatment, but the observer can give feedback on the decision maker’s choice. Like in the Decision and Photo treatments, the observer can select either a “like” or a “dislike” button.

2024-02-01

2024-03-14

Amount of money sent in the dictator game

Beliefs on social norms, social image concerns, Other in the Self (IOS) Scale, behavioral change in the amount sent in the dictator game between Part 1 and Part 2 (frequency and magnitude)

We will conduct an incentivized between-subject online experiment aimed at untangling possible relevant mechanisms that drive the impact of observability on pro-social behavior. Specifically, we will examine two key factors: the level of observability (knowing there will be an observer; exchanging photos; seeing each other in a video call) and feedback on the decision (like/dislike).

The experiment consists of two parts and a survey. In the first part, depending on the treatment, participants play a dictator game with an observer. Thus, there are three different roles: decision maker, receiver, and observer. Participants remain in the same role throughout the experiments. We will implement six experimental conditions and a control condition:

• Control condition: Participants will play a standard dictator game. There is no observer in this condition.
• Decision Treatment: The decision maker’s choice of the amount sent to an anonymous receiver is shared with an anonymous observer.
• Decision plus Feedback Treatment: Like the Decision treatment, but the observer can give feedback on the decision maker’s choice. Specifically, the observer can select either a “like” or a “dislike” button.
• Photo Treatment: Like the Decision treatment, but additionally the decision maker and the observer exchange photos after the decision has been made.
• Photo plus Feedback Treatment: Like the Photo treatment, but the observer can give feedback on the decision maker’s choice. Like in the decision treatment, the observer can select either a “like” or a “dislike” button.
• Video Treatment: Like the Photo treatment, but instead of exchanging photos, decision maker and observer meet in a video call where they can see each other but cannot communicate verbally.
• Video plus Feedback Treatment: Like the Video treatment, but the observer can give feedback on the decision maker’s choice. Like in the Decision and Photo treatments, the observer can select either a “like” or a “dislike” button.

In the second part, independently of the treatment, participants play a standard dictator game with no observer. They are randomly matched with a new participant in the role of the receiver. This allows us to analyze behavioral stickiness or change after observation vanishes.

It is common knowledge that at the end of the study, it will be decided randomly, with equal probability, whether Part 1 or Part 2 will be relevant for payment. It is also common knowledge that the participant in the role of the receiver is recruited after the decision makers have made their decisions.

In addition, we will elicit decision makers’ beliefs on descriptive and injunctive social norms. Belief elicitation will be incentivized and takes place after participants made their decision in Part 1 of the experiment. Observers are asked to indicate the minimum sending amount they would like/dislike.

Participants will also complete a short survey on self-report scales (e.g., social image concerns, Inclusion of the Other in the Self (IOS) Scale) and answer demographic questions.

Not available

online by a computer

individual

No

1470

The experiment will be run on Prolific. The number of participants planned for the experiment is 4200 people completing the study. We advertise for 4730 participants to take into account individuals who might not be matched and/or who must be excluded due to the restrictions below. We will exclude subjects who: - complete the survey faster than two standard deviations from the average completion time; - do not complete the study within 30 minutes of starting; - do not complete the study for other reasons (time runs out, no match in the defined time) - exit and then re-enter the task as a new subject (as these individuals might see multiple treatments); - are not approved for any other reason (e.g., not having a valid Prolific ID, not having a working webcam); - do not answer the control question correctly the second time - are not aware that they and their decisions are observed (if this is the case in the corresponding treatment) - whose face is not visible in the call or on the photo Based on previous research, we expect to exclude about 5-10 percent of participants because they do not meet the restrictions above.

For each of the experimental treatments: About 210 groups consisting of three participants (1 decision maker, 1 observer, 1 receiver)
For the control treatment: about 210 groups consisting of two participants (1 decision maker, 1 receiver)

Based on a two-sided Wilcoxon-Mann-Whitney test, an error probability of 0.05, and a power of 0.80, we require about 210 groups per treatment to detect an effect of Cohen’s d of 0.28.