Gender Interaction in Teams: Experimental Evidence on Performance and Punishment Behavior

Abstract

This paper reports results from a real-effort experiment in which men and women are paired to form a two-member team and asked to execute a real-effort task. Each participant receives an equal share of the team's output. Workers who perform better than their partner can punish him/her by imposing a fine. We manipulate the teams' gender composition (man-man, man-woman, and woman-woman) to analyze whether an individual's performance and sanctioning behavior depends on his/her gender and the gender interaction within the team. The data show that, on average, men perform slightly better than women. A man's performance will deteriorate when paired with a woman, while a woman's performance will improve when paired with a woman. When underperforming, women are sanctioned more often and more heavily than men; if sanctioned, men tend to improve their performance, while women's performance does not change.

Experimental Evidence on Gender Interaction in Lying Behavior

“…It was also important to ensure that team members were aware of the gender of their partner, while preserving anonymity and without conveying this information in a salient way. We thus followed the method used by Jung and Vranceanu (2015) in organizing sessions where only men (only women) were invited, and mixed-gender sessions with an equal number of men and women. In the single-gender sessions participants can clearly observe that only men (women) are present, so they realized that they were paired with a same-gender partner.…”

“…As it has been argued by Besancenot et al (2013) by means of a simple model, if (1) Responders have fairness concerns and (2) there are at least some Proposers who always tell the truth (have very high lying costs), it is in the interest of the other Proposers to understate the amount received because Bayesian Responders would (partially) trust them, and would therefore accept lower offers. 4 Our methodology builds on the experimental design by Besancenot et al (2013), with the notable difference that subjects are now aware of the gender of their (anonymous) partner (Jung and Vranceanu (2015)). In a nutshell, at the outset of typical round, the computer draws the Proposer's endowment form the set of integer numbers [50;100].…”

Experimental Evidence on Gender Interaction in Lying Behavior

“…As it has been argued by Besancenot et al (2013) by means of a simple model, if (1) Responders have fairness concerns and (2) there are at least some Proposers who always tell the truth (have very high lying costs), it is in the interest of the other Proposers to understate the amount received because Bayesian Responders would (partially) trust them, and would therefore accept lower offers. 4 Our methodology builds on the experimental design by Besancenot et al (2013), with the notable difference that subjects are now aware of the gender of their (anonymous) partner (Jung and Vranceanu (2015)). In a nutshell, at the outset of typical round, the computer draws the Proposer's endowment form the set of integer numbers [50;100].…”

Willingness to compete: Between‐ and within‐gender comparisons

“…This method to release gender information in a “smooth” way has already been used by Jung and Vranceanu (2017b,a) in a different experiment.…”