Opening the black box of emotion expression: How anger expression shapes beliefs, attitudes and tastes in negotiation

Description of the Experiment
Subjects are randomly assigned to one of four conditions in a 2×2 design that varies the expression in a verbal message (neutra l vs. anger) and the counterpart’s gender (female vs. male). In all conditions, participants make decisions in an ultimatum game and a dictator game after viewing a treatment-specific message from their counterpart. They also make choices between lottery pairs to elicit beliefs and ambiguity attitudes.

This design allows us to cleanly identify the causal effect of anger expression in negotiation and to uncover the mechanisms behind it. Specifically, we focus on three questions: 1) Do individuals make larger concessions in negotiations when their counterparts express anger?
2) Are any such differences driven by changes in altruism, beliefs, or ambiguity attitudes? 3)Does the effect of anger depend on the counterpart’s gender?

2025-12-01

2026-04-01

- Ultimatum game offer
- Ambiguity aversion index
- A-insensitivity index
- Expected reservation price
- Dictator game offer

We elicit participants’ ambiguity aversion index, a-insentivitiy index and expected reservation price using the belief-hedge method (Baillon et al., 2018, 2021). From participants’ choices between ambiguous and risky options, we obtain matching probabilities that allow us to construct these three primary outcomes.

We adopt a 2*2 between-subject design: anger vs neutral verbal message and male vs female counterpart. Participants and counterparts remain anonymous, and gender is revealed only through a gender-specific pseudonym and the avatar presented in front of the message box.

In the experiment, participants act as proposers in negotiation games. They first answer a few demographic questions and are then randomly and anonymously paired with another participant who serves as their counterpart (female or male).

The first task is an Ultimatum Game. Each participant receives an endowment of $20 and decides how much of it to offer to their counterpart. If the counterpart accepts the offer, the participant receives the remaining amount from the $20 and the counterpart receives the offered amount. If the counterpart rejects the offer, both receive $0. Before participants decide on their offer, they receive a verbal message from the counterpart, which is either angry or neutral.

The second task is a bi-section choice task, based on the belief-hedge method introduced by Baillon et al. (2018, 2021). Participants make decisions between ambiguous and risky options. The probability in the risky option is varied until indifference is reached within a pre-determined error margin. These choice situations are designed to elicit the matching probability associated with events concerning the counterpart’s reservation price. The resulting matching probabilities allow us to construct three indexes: (i) an ambiguity-aversion index, capturing aversion to ambiguity about the counterpart’s reservation price; (ii) an ambiguity-insensitivity index, capturing how precisely or imprecisely participants perceive this uncertainty; and (iii) an expected reservation price, reflecting participants’ beliefs about what their counterpart is likely to accept.

The third task is a Dictator Game. Each participant is paired with a different counterpart and receives an endowment of $20. Participants decide how much of this amount to give to the counterpart. The counterpart has no decision power and simply receives whatever amount is allocated. Before making their allocation decision, participants again receive a verbal message from the counterpart. The counterpart’s gender and the emotion expression of the message are the same as in the Ultimatum Game.

At the end of the experiment, we elicit participants’ attachment styles and collect additional demographic information.

References:
Baillon, A., Bleichrodt, H., Li, C., & Wakker, P. P. (2021). Belief hedges: Measuring ambiguity for all events and all models. Journal of Economic Theory, 198, 105353.
Baillon, A., Huang, Z., Selim, A., & Wakker, P. P. (2018). Measuring ambiguity attitudes for all (natural) events. Econometrica, 86(5), 1839-1858.

Not available

Randomization is implemented automatically through Qualtrics’ built-in randomizer.

individual

No

No clustering

800 online participants

We have four treatments (female counterpart x anger message; female counterpart x neutral message; male counterpart x anger message; male counterpart x neutral message). We will have 200 online participants per treatment.