Gender Differences in Multilateral Bargaining

In a laboratory experiment, we exogenously manipulate the gender composition of the group in a between subjects design to test, whether gender differences appear in a simple three player bargaining game, when the information about the gender composition of the group is revealed to the subjects. The experiment entails 4 treatments: all female, all male, female majority and male majority groups. A control group could be considered the all male treatment with an interest of changes in outcomes as more females are added to the game.

2019-02-18

2019-02-21

We will observe the effects of the gender composition of the group on the following outcomes: endogenously arising characteristics of the proposer, proportion of formed minimal winning coalitions, gender composition of the resulting coalitions, efficiency and equality of bargaining outcomes, as well as individual performance.

To describe minimal winning coalition formation, we will look at the proportion of agreements that assign positive payoffs to two out of three group members. Gender composition of the coalitions will be indicated by the proportions of male, female and mixed gender coalitions. To describe efficiency, we will look at the average time, in seconds, that it takes for a group to reach an agreement, cumulative distribution of agreements across time, as well as the proportion of the surplus acquired. The equality of outcomes will be described by the distribution of splits in each treatment. We will measure individual performance (including proposer's advantage) by looking at the likelihood of inclusion into a minimal winning coalition, average absolute as well as proportional payoffs, the difference between the individually acquired share and the largest share of another subject in the same group. Finally, we will observe the gender and timing of the initial proposer.

Secondary outcomes are obtained from the survey module responses at the end of the experiment. We will collect data on individual risk aversion, time discounting, trust, altruism, positive and negative reciprocity (Falk et al, 2018), as well as competitiveness (Reuben et al), and scores in the cognitive reflection test. We will also observe subject responses to the questions about the purpose of the experiment and salience of the information about the gender composition of the group.

3 symmetric agents make costless proposals in an unstructured setting to bargain over a decaying surplus. Allocations are confirmed through a majority voting rule. Interactions are repeated for 12 rounds with random rematching. At the end of the experiment, subjects complete a survey module to provide responses for the exploratory analysis of the main experimental data.

3 symmetric agents make costless proposals in an unstructured setting to bargain over a decaying surplus for 150 seconds. The surplus is 12 points that individuals can allocate among themselves. Initially, each point is worth €3. Thus, initially, subjects have €3× 12 points = €36 to divide within their group. The exchange rate of €3 declines at a rate of €0.02 per second. We implement majority decision rule with a ratification period of 10 seconds. That is, bargaining ends, when an agreement is reached (when at least two participants continuously agree on the allocation of points for 10 seconds) or when the exchange rate reaches €0 per point. If there is no agreement after the exchange rate has reached €0 per point, bargaining ends, and no one receives any points. If an agreement is finalized, potential subject earnings are determined by the number of points allocated to him/her and the exchange rate at the time when the allocation is confirmed (when 10 second ratification period expires).

At the beginning of the experiment, subjects complete a short demographics questionnaire that includes a question about their sex. They receive instructions about the bargaining game and complete 20 minute tutorial to learn to use the computer interface to make decisions. The above bargaining game is repeated 12 times (subjects participate in 12 rounds of the bargaining interactions). In each round, subjects are randomly rematched within their matching group. The feedback between rounds includes information on whether an agreement has been reached, the number of points allocated to the subject, the exchange rate at which the allocation has been confirmed.

At the end of the 12 rounds, subjects answer questions of a survey module for measuring risk aversion, time discounting, trust, altruism, positive and negative reciprocity (Falk et al, 2018), as well as competitiveness (Reuben et al). They complete a cognitive reflection test. We include questions about the purpose of the experiment and salience of the information about the gender composition of the group.

One out of twelve rounds is randomly selected for payment. Expected per subject payment is €17 (including €5 show up payment). Experiment is coded in z-Tree (Fischbacher, 2007) and subjects are recruited using ORSEE (Greiner, 2015) at BEElab, Maastricht University.

We are planning to run 12 sessions in the week of February 18-22, 2019. Each session consists of either all female matching group of 12 subjects AND all male matching group of 12 subjects, OR female majority matching group of 12 subjects AND male majority matching group of 12 subjects. In such a way, our sessions are always gender balanced. This feature may be useful, if we decide to run additional treatments with no information about the gender composition of the group.

Randomization done in the lab by a computer.

Individuals are randomly assigned to matching groups of 12 subjects within 24 subject sessions. They play 12 rounds of a bargaining game with random rematching within their matching group. Each treatment consists of 6 matching groups.

Yes

24 matching groups.

288 participants.

6 matching groups (72 subjects) per treatment.